research abstract samples

Purdue Online Writing Lab Purdue OWL® College of Liberal Arts

Welcome to the Purdue OWL

This page is brought to you by the OWL at Purdue University. When printing this page, you must include the entire legal notice.

Copyright ©1995-2018 by The Writing Lab & The OWL at Purdue and Purdue University. All rights reserved. This material may not be published, reproduced, broadcast, rewritten, or redistributed without permission. Use of this site constitutes acceptance of our terms and conditions of fair use.

Scholars often write abstracts for various applications: conference presentations may require an abstract or other short summary for a program; journal articles almost always require abstracts; invited talks and lectures are often advertised using an abstract. While the application may necessarily change the length of the abstract (a conference program may only allow for 50-75 words, for instance), the purpose and structure remains fairly constant.

Abstracts are generally kept brief (approximately 150-200 words). They differ by field, but in general, they need to summarize the article so that readers can decide if it is relevant to their work. The typical abstract includes these elements:

• A statement of the problem and objectives
• A statement of the significance of the work
• A summary of employed methods or your research approach
• A summary of findings or conclusions of the study
• A description of the implications of the findings

Regardless of field, abstract authors should explain the purpose of the work, methods used, the results and the conclusions that can be drawn. However, each field purports slightly different ways to structure the abstract. A reliable strategy is to write the abstract as a condensed version of your article, with 1-2 sentences summarizing each major section. This means that in many of the sciences and a large portion of the humanities, abstracts follow a version of the IMRAD structure: Introduction, Methods, Results, and Discussion.

Most scientific journals require authors to submit such abstracts. It is generally advisable to write the abstract in the English language. That is because most papers in other languages, especially Asian nations, tend to publish an English abstract with common search engines, such as, the MLA site.

Example Abstract

This example abstract follows the IMRAD structure closely. The first two sentences are the introduction and background information. Sentences 3-5 describe the methods used in the study. Sentence 6 summarizes the results, while the last two sentences summarize the discussion and conclusion of the study; they also indicate the significance of the results.

Usability and User-Centered Theory for 21 st Century OWLs — by Dana Lynn Driscoll, H. Allen Brizee, Michael Salvo, and Morgan Sousa from The Handbook of Research on Virtual Workplaces and the New Nature of Business Practices . Eds. Kirk St. Amant and Pavel Zemlansky. Hershey, PA: Idea Group Publishing, 2008.

This article describes results of usability research conducted on the Purdue Online Writing Lab (OWL). The Purdue OWL is an information-rich educational website that provides free writing resources to users worldwide. Researchers conducted two generations of usability tests. In the first test, participants were asked to navigate the OWL and answer questions. Results of the first test and user-centered scholarship indicated that a more user-centered focus would improve usability. The second test asked participants to answer writing-related questions using both the OWL website and a user-centered OWL prototype. Participants took significantly less time to find information using the prototype and reported a more positive response to the user-centered prototype than the original OWL. Researchers conclude that a user-centered website is more effective and can be a model for information-rich online resources. Researchers also conclude that usability research can be a productive source of ideas, underscoring the need for participatory invention.

Academic & Employability Skills

Subscribe to academic & employability skills.

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 408 other subscribers.

Email Address

Writing an abstract - a six point checklist (with samples)

Posted in: abstract , dissertations

The abstract is a vital part of any research paper. It is the shop front for your work, and the first stop for your reader. It should provide a clear and succinct summary of your study, and encourage your readers to read more. An effective abstract, therefore should answer the following questions:

• Why did you do this study or project?
• What did you do and how?
• What did you find?
• What do your findings mean?

So here's our run down of the key elements of a well-written abstract.

• Size - A succinct and well written abstract should be between approximately 100- 250 words.
• Background - An effective abstract usually includes some scene-setting information which might include what is already known about the subject, related to the paper in question (a few short sentences).
• Purpose  - The abstract should also set out the purpose of your research, in other words, what is not known about the subject and hence what the study intended to examine (or what the paper seeks to present).
• Methods - The methods section should contain enough information to enable the reader to understand what was done, and how. It should include brief details of the research design, sample size, duration of study, and so on.
• Results - The results section is the most important part of the abstract. This is because readers who skim an abstract do so to learn about the findings of the study. The results section should therefore contain as much detail about the findings as the journal word count permits.
• Conclusion - This section should contain the most important take-home message of the study, expressed in a few precisely worded sentences. Usually, the finding highlighted here relates to the primary outcomes of the study. However, other important or unexpected findings should also be mentioned. It is also customary, but not essential, to express an opinion about the theoretical or practical implications of the findings, or the importance of their findings for the field. Thus, the conclusions may contain three elements:
• The primary take-home message.
• Any additional findings of importance.
• Implications for future studies.

Example Abstract 2: Engineering Development and validation of a three-dimensional finite element model of the pelvic bone.

Abstract from: Dalstra, M., Huiskes, R. and Van Erning, L., 1995. Development and validation of a three-dimensional finite element model of the pelvic bone. Journal of biomechanical engineering, 117(3), pp.272-278.

And finally...  A word on abstract types and styles

Abstract types can differ according to subject discipline. You need to determine therefore which type of abstract you should include with your paper. Here are two of the most common types with examples.

Informative Abstract

The majority of abstracts are informative. While they still do not critique or evaluate a work, they do more than describe it. A good informative abstract acts as a surrogate for the work itself. That is, the researcher presents and explains all the main arguments and the important results and evidence in the paper. An informative abstract includes the information that can be found in a descriptive abstract [purpose, methods, scope] but it also includes the results and conclusions of the research and the recommendations of the author. The length varies according to discipline, but an informative abstract is usually no more than 300 words in length.

Descriptive Abstract A descriptive abstract indicates the type of information found in the work. It makes no judgements about the work, nor does it provide results or conclusions of the research. It does incorporate key words found in the text and may include the purpose, methods, and scope of the research. Essentially, the descriptive abstract only describes the work being summarised. Some researchers consider it an outline of the work, rather than a summary. Descriptive abstracts are usually very short, 100 words or less.

Adapted from Andrade C. How to write a good abstract for a scientific paper or conference presentation. Indian J Psychiatry. 2011 Apr;53(2):172-5. doi: 10.4103/0019-5545.82558. PMID: 21772657; PMCID: PMC3136027 .

Share this:

• Click to print (Opens in new window)
• Click to email a link to a friend (Opens in new window)
• Click to share on WhatsApp (Opens in new window)
• Click to share on LinkedIn (Opens in new window)
• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Click here to cancel reply.

• Email * (we won't publish this)

Write a response

Navigating the dissertation process: my tips for final years

Imagine for a moment... After months of hard work and research on a topic you're passionate about, the time has finally come to click the 'Submit' button on your dissertation. You've just completed your longest project to date as part...

8 ways to beat procrastination

Whether you’re writing an assignment or revising for exams, getting started can be hard. Fortunately, there’s lots you can do to turn procrastination into action.

My takeaways on how to write a scientific report

If you’re in your dissertation writing stage or your course includes writing a lot of scientific reports, but you don’t quite know where and how to start, the Skills Centre can help you get started. I recently attended their ‘How...

• Bipolar Disorder
• Therapy Center
• When To See a Therapist
• Types of Therapy
• Best Online Therapy
• Best Couples Therapy
• Best Family Therapy
• Managing Stress
• Sleep and Dreaming
• Understanding Emotions
• Self-Improvement
• Healthy Relationships
• Student Resources
• Personality Types
• Guided Meditations
• Verywell Mind Insights
• 2024 Verywell Mind 25
• Mental Health in the Classroom
• Editorial Process
• Meet Our Review Board
• Crisis Support

How to Write an APA Abstract

Kendra Cherry, MS, is a psychosocial rehabilitation specialist, psychology educator, and author of the "Everything Psychology Book."

Emily is a board-certified science editor who has worked with top digital publishing brands like Voices for Biodiversity, Study.com, GoodTherapy, Vox, and Verywell.

Verywell / Nusha Ashjaee 

• Writing Your Abstract
• How to Use Keywords

An APA abstract is a concise but comprehensive summary of a scientific paper. It is typically a paragraph long, or about 150 to 250 words. The goal of the abstract is to provide the reader with a brief and accurate idea of what a paper is about.

The APA abstract should appear on a separate page immediately after the title page and before the main content of your paper. While professional papers that appear in scientific journals and other publications require an APA abstract, they may not be required for student papers. However, you should always check with your instructor for specific requirements.

What Is APA Format?

APA format is the official style of the American Psychological Association. It is used in writing for psychology and other social sciences. These style guidelines specify different aspects of a document's presentation and layout, including how pages are structured, how references are organized, and how sources are cited.

This article explains how to create an abstract in APA format for your psychology papers or other types of scientific writing. It covers the basic rules you should follow as well as specific guidelines for writing abstracts for experimental reports, literature reviews, and other articles.

What Is an Abstract in APA Format?

In addition to providing guidance for the general style and organization of a paper, APA format also stipulates using an abstract designed to briefly summarize the key details in a paper.

While it is sometimes overlooked or only an afterthought, an abstract is an integral part of any academic or professional paper. The abstract is a critical component of an APA-formatted paper. This brief overview summarizes what your paper contains. It should succinctly and accurately represent what your paper is about and what the reader can expect to find.

Following a few simple guidelines, you can create an abstract following the format. Done well, an abstract generates interest in your work and helps readers learn if the paper will interest them.

APA Format Abstract Basics

The abstract is the second page of a lab report or APA-format paper and should immediately follow the title page . Think of an abstract as a highly condensed summary of your entire paper.

The purpose of your abstract is to provide a brief yet thorough overview of your paper. It should function much like your title page—it should allow the person reading it to quickly determine what your paper is all about. Your abstract is the first thing that most people will read, and it is usually what informs their decision to read the rest of your paper.

The abstract is the single most important paragraph in your entire paper, according to the APA Publication Manual. A good abstract lets the reader know that your paper is worth reading.

According to the official guidelines of the American Psychological Association, an abstract should be brief but packed with information. Each sentence must be written with maximum impact in mind. To keep your abstract short, focus on including just four or five of the essential points, concepts, or findings.

An abstract must also be objective and accurate. The abstract's purpose is to report rather than provide commentary. It should accurately reflect what your paper is about. Only include information that is also included in the body of your paper.

Key Elements of an APA Abstract

Your abstract page should include:

• A running head , which is a shortened version of your title that appears in all caps at the top left of each page of your paper
• A section label , which should be the word "Abstract" centered and bolded at the top of the page
• A page number , which should be the second page of your paper (the title page should be page 1)
• A double-spaced paragraph of about 150 to 250 words
• An indented list of keywords related to your paper's content. Include the label "Keywords:" in italics and list three to five keywords that are separated by commas

How to Write an Abstract in APA Format

Before you write your abstract, you first need to write your paper in its entirety. In order to write a good abstract, you need to have a finished draft of your paper so you can summarize it accurately.

While the abstract will be at the beginning of your paper, it should be the last section you write.

Once you have completed the final draft of your psychology paper , use it as a guide for writing your abstract.

• Begin your abstract on a new page . Place your running head and page number 2 in the top right-hand corner. Center the word "Abstract" at the top of the page.
• Know your target word count . An abstract should be between 150 and 250 words. Exact word counts vary from journal to journal . If you are writing your paper for a psychology course, your professor may have specific word requirements, so be sure to ask. The abstract should be written as only one paragraph with no indentation.
• Structure the abstract in the same order as your paper . Begin with a brief summary of the introduction , and then continue on with a summary of the method , results , and discussion sections of your paper.
• Look at other abstracts in professional journals for examples of how to summarize your paper . Notice the main points that the authors chose to mention in the abstract. Use these examples as a guide when choosing the main ideas in your own paper.
• Write a rough draft of your abstract . Use the format required for your type of paper (see next sections). While you should aim for brevity, be careful not to make your summary too short. Try to write one to two sentences summarizing each section of your paper. Once you have a rough draft, you can edit for length and clarity.
• Ask a friend to read over the abstract . Sometimes, having someone look at your abstract with fresh eyes can provide perspective and help you spot possible typos and other errors.

The abstract is vital to your paper, so it should not be overlooked or treated as an afterthought. Spend time writing this section carefully to ensure maximum readability and clarity.

It is important to remember that while the abstract is the last thing you write, it is often the most read part of your paper.

Experimental Report Abstracts

The format of your abstract also depends on the type of paper you are writing. For example, an abstract summarizing an experimental paper will differ from that of a meta-analysis or case study . For an experimental report, your abstract should:

• Identify the problem . In many cases, you should begin by stating the question you sought to investigate and your hypothesis .
• Describe the participants in the study . State how many participants took part and how they were selected. For example: "In this study, 215 undergraduate student participants were randomly assigned to [the experimental condition] or [the control condition]."
• Describe the study method . For example, identify whether you used a within-subjects, between-subjects, or mixed design.
• Give the basic findings . This is essentially a brief preview of the results of your paper. 
• Provide any conclusions or implications of the study . What might your results indicate, and what directions does it point to for future research?

Literature Review Abstracts

If your paper is a meta-analysis or literature review, your abstract should:

• Describe the problem of interest . In other words, what is it that you set out to investigate in your analysis or review?
• Explain the criteria used to select the studies included in the paper . There may be many different studies devoted to your topic. Your analysis or review probably only looks at a portion of these studies. For what reason did you select these specific studies to include in your research?
• Identify the participants in the studies . Inform the reader about who the participants were in the studies. Were they college students? Older adults? How were they selected and assigned?
• Provide the main results . Again, this is essentially a quick peek at what readers will find when they read your results section. Don't try to include everything. Just provide a very brief summary of your main findings. 
• Describe any conclusions or implications . What might these results mean and what do they reveal about the body of research that exists on this particular topic?

Lab Reports and Articles

Psychology papers such as lab reports and APA format articles also often require an abstract. In these cases as well, the abstract should include all of the major elements of your paper, including an introduction, hypothesis, methods, results, and discussion.

Remember, although the abstract should be placed at the beginning of your paper (right after the title page), you will write the abstract last after you have completed a final draft of your paper.

To ensure that all of your APA formatting is correct, consider consulting a copy of the  Publication Manual of the American Psychological Association .

Keywords in an APA Abstract

After the paragraph containing the main elements of your abstract, you can also include keywords related to your paper. Such keywords are used when indexing your paper in databases and can help researchers and students locate your paper when searching for information about those topics.

Because keywords help people find your paper, it is essential to choose the right ones. The APA suggests including between three and five keywords.

You can identify keywords by thinking about what your paper is about. For example, if your paper focuses on how social media use is related to depression in teenagers, you might include the keywords: social media, mood, depression, adolescents, social networking sites 

A Word From Verywell

The abstract may be very brief, but it is so important that the official APA style manual identifies it as the most important paragraph in your entire paper. Careful attention to detail can ensure that your abstract does a good job representing the contents of your paper. If possible, take your paper to your school's writing lab for assistance.

Nagda S. How to write a scientific abstract. J Indian Prosthodont Soc. 2013;13(3):382–383. doi:10.1007/s13191-013-0299-x

Kumar A. Writing an abstract: Revealing the essence with eloquence .  J Indian Soc Periodontol . 2022;26(1):1-2. doi:10.4103/jisp.jisp_634_21

American Psychological Association. APA Style Journal Article Reporting Standards: Reporting Standards for Studies With an Experimental Manipulation .

American Psychological Association. APA Style Journal Article Reporting Standards: Quantitative Meta-Analysis Article Reporting Standards .

Tullu MS. Writing the title and abstract for a research paper: Being concise, precise, and meticulous is the key .  Saudi J Anaesth . 2019;13(Suppl 1):S12-S17. doi:10.4103/sja.SJA_685_18

American Psychological Association. Publication Manual of the American Psychological Association (7th ed.). American Psychological Association; 2019.

By Kendra Cherry, MSEd Kendra Cherry, MS, is a psychosocial rehabilitation specialist, psychology educator, and author of the "Everything Psychology Book."

Journal Article: Abstract

When to write the abstract.

• Introduction

Writing an abstract can be difficult because you are tasked with condensing tons of work into such a small amount of space. To make things easier, write your abstract last. Read through your entire paper and distill each section down to its main points. Sometimes it can be helpful to answer this question through a subtractive process. For example, if you are trying to distill down your results, simply list all your findings and then go through that list and start crossing off or consolidating each finding until you are left with a only the most crucial results.

Your title and abstract are the primary medium through which interested readers will find your work amidst the deluge of scientific publications, posters, or conference talks. When a fellow scientist happens upon your abstract they will quickly skim it to determine if it is worth their time to dive into the main body of the paper. The main purpose of an abstract, therefore, is to contextualize and describe your work in a concise and easily-understood manner. This will ensure that your scientific work is found and read by your intended audience.

Abstract Formula

Clarity is achieved by providing information in a predictable order: successful abstracts therefore are composed of 6 ordered components which are referred to as the “abstract formula”.

General and   Specific Background (~1 sentence each). Introduce the area of science that you will be speaking about and the state of knowledge in that area. Start broad in the general background, then narrow in on the relevant topic that will be pursued in the paper. I f you use jargon, be sure to very briefly define it.

Knowledge Gap (~1 sentence). Now that you’ve stated what is already known, state what is not known. W hat specific question is your work attempting to answer?

“Here we show…” (~1 sentence). State your general experimental approach and the answer to the question which you just posed in the “Knowledge Gap” section.

Experimental Approach & Results (~1-3 sentences). Provide a high-level description of your most important methods and results. How did you get to the conclusion that you stated in the “Here we show…” section?

Implications (~1 sentence).  Describe how your findings influence our understanding of the relevant field and/or their implications for future studies.

This content was adapted from from an article originally created by the  MIT Biological Engineering Communication Lab .

Resources and Annotated Examples

Annotated example 1.

Annotated abstract of a microbiology paper published in Nature . 4 MB

Annotated Example 2

Annotated abstract of a paper published in Nature . 2 MB

How to Write an Abstract for a Scientific Paper

• Chemical Laws
• Periodic Table
• Projects & Experiments
• Scientific Method
• Biochemistry
• Physical Chemistry
• Medical Chemistry
• Chemistry In Everyday Life
• Famous Chemists
• Activities for Kids
• Abbreviations & Acronyms
• Weather & Climate
• Ph.D., Biomedical Sciences, University of Tennessee at Knoxville
• B.A., Physics and Mathematics, Hastings College

If you're preparing a research paper or grant proposal, you'll need to know how to write an abstract. Here's a look at what an abstract is and how to write one.

An abstract is a concise summary of an experiment or research project. It should be brief -- typically under 200 words. The purpose of the abstract is to summarize the research paper by stating the purpose of the research, the experimental method, the findings, and the conclusions.

• How to Write an Abstract

The format you'll use for the abstract depends on its purpose. If you're writing for a specific publication or a class assignment, you'll probably need to follow specific guidelines. If there isn't a required format, you'll need to choose from one of two possible types of abstracts.

Informational Abstracts

An informational abstract is a type of abstract used to communicate an experiment or lab report .

• An informational abstract is like a mini-paper. Its length ranges from a paragraph to 1 to 2 pages, depending on the scope of the report. Aim for less than 10% the length of the full report.
• Summarize all aspects of the report, including purpose, method, results, conclusions, and recommendations. There are no graphs, charts, tables, or images in an abstract. Similarly, an abstract does not include a bibliography or references.
• Highlight important discoveries or anomalies. It's okay if the experiment did not go as planned and necessary to state the outcome in the abstract.

Here is a good format to follow, in order, when writing an informational abstract. Each section is a sentence or two long:

• Motivation or Purpose: State why the subject is important or why anyone should care about the experiment and its results.
• Problem: State the hypothesis of the experiment or describe the problem you are trying to solve.
• Method: How did you test the hypothesis or try to solve the problem?
• Results: What was the outcome of the study? Did you support or reject a hypothesis? Did you solve a problem? How close were the results to what you expected? State-specific numbers.
• Conclusions: What is the significance of your findings? Do the results lead to an increase in knowledge, a solution that may be applied to other problems, etc.?

Need examples? The abstracts at PubMed.gov (National Institutes of Health database) are informational abstracts. A random example is this abstract on the effect of coffee consumption on Acute Coronary Syndrome .

Descriptive Abstracts

A descriptive abstract is an extremely brief description of the contents of a report. Its purpose is to tell the reader what to expect from the full paper.

• A descriptive abstract is very short, typically less than 100 words.
• Tells the reader what the report contains, but doesn't go into detail.
• It briefly summarizes the purpose and experimental method, but not the results or conclusions. Basically, say why and how the study was made, but don't go into findings. 

Tips for Writing a Good Abstract

• Write the paper before writing the abstract. You might be tempted to start with the abstract since it comes between the title page and the paper, but it's much easier to summarize a paper or report after it has been completed.
• Write in the third person. Replace phrases like "I found" or "we examined" with phrases like "it was determined" or "this paper provides" or "the investigators found".
• Write the abstract and then pare it down to meet the word limit. In some cases, a long abstract will result in automatic rejection for publication or a grade!
• Think of keywords and phrases a person looking for your work might use or enter into a search engine. Include those words in your abstract. Even if the paper won't be published, this is a good habit to develop.
• All information in the abstract must be covered in the body of the paper. Don't put a fact in the abstract that isn't described in the report.
• Proof-read the abstract for typos, spelling mistakes, and punctuation errors.
• Abstract Writing for Sociology
• How to Write a Science Fair Project Report
• How to Format a Biology Lab Report
• Six Steps of the Scientific Method
• How to Write a Lab Report
• Science Lab Report Template - Fill in the Blanks
• Scientific Method Flow Chart
• What Is Depth of Knowledge?
• How To Design a Science Fair Experiment
• An Introduction to Academic Writing
• Null Hypothesis Definition and Examples
• How to Write a Research Paper That Earns an A
• 6 Steps to Writing the Perfect Personal Essay
• Writing a Lesson Plan: Closure and Context

Abstracts are formal summaries writers prepare of their completed work. Abstracts are important tools for readers, especially as they try to keep up with an explosion of information in print and on the Internet.

Definition of Abstract

Abstracts, like all summaries, cover the main points of a piece of writing. Unlike executive summaries written for non-specialist audiences, abstracts use the same level of technical language and expertise found in the article itself. And unlike general summaries which can be adapted in many ways to meet various readers' and writers' needs, abstracts are typically 150 to 250 words and follow set patterns.

Because readers use abstracts for set purposes, these purposes further define abstracts.

Purposes for Abstracts

Abstracts typically serve five main goals:

Help readers decide if they should read an entire article

Readers use abstracts to see if a piece of writing interests them or relates to a topic they're working on. Rather than tracking down hundreds of articles, readers rely on abstracts to decide quickly if an article is pertinent. Equally important, readers use abstracts to help them gauge the sophistication or complexity of a piece of writing. If the abstract is too technical or too simplistic, readers know that the article will also be too technical or too simplistic.

Help readers and researchers remember key findings on a topic

Even after reading an article, readers often keep abstracts to remind them of which sources support conclusions. Because abstracts include complete bibliographic citations, they are helpful when readers begin writing up their research and citing sources.

Help readers understand a text by acting as a pre-reading outline of key points

Like other pre-reading strategies, reading an abstract before reading an article helps readers anticipate what's coming in the text itself. Using an abstract to get an overview of the text makes reading the text easier and more efficient.

Index articles for quick recovery and cross-referencing

Even before computers made indexing easier, abstracts helped librarians and researchers find information more easily. With so many indexes now available electronically, abstracts with their keywords are even more important because readers can review hundreds of abstracts quickly to find the ones most useful for their research. Moreover, cross-referencing through abstracts opens up new areas of research that readers might not have known about when they started researching a topic.

Allow supervisors to review technical work without becoming bogged down in details

Although many managers and supervisors will prefer the less technical executive summary, some managers need to keep abreast of technical work. Research shows that only 15% of managers read the complete text of reports or articles. Most managers, then, rely on the executive summary or abstract as the clearest overview of employees' work.

Types of Abstracts

Although you'll see two types of abstracts—informative and descriptive—most writers now provide informative abstracts of their work.

Descriptive Abstract

A descriptive abstract outlines the topics covered in a piece of writing so the reader can decide whether to read the entire document. In many ways, the descriptive abstract is like a table of contents in paragraph form. Unlike reading an informative abstract, reading a descriptive abstract cannot substitute for reading the document because it does not capture the content of the piece. Nor does a descriptive abstract fulfill the other main goals of abstracts as well as informative abstracts do. For all these reasons, descriptive abstracts are less and less common. Check with your instructor or the editor of the journal to which you are submitting a paper for details on the appropriate type of abstract for your audience.

Informative Abstract

An informative abstract provides detail about the substance of a piece of writing because readers will sometimes rely on the abstract alone for information. Informative abstracts typically follow this format:

• Identifying information (bibliographic citation or other identification of the document)
• Concise restatement of the main point, including the initial problem or other background
• Methodology (for experimental work) and key findings
• Major conclusions

Informative abstracts usually appear in indexes like Dissertation Abstracts International ; however, your instructor may ask you to write one as a cover sheet to a paper as well.

A More Detailed Comparison of Descriptive vs. Informative

The typical distinction between descriptive and informative is that the descriptive abstract is like a table of contents whereas the informative abstract lays out the content of the document. To show the differences as clearly as possible, we compare a shortened Table of Contents for a 100-page legal argument presented by the FDA and an informative abstract of the judge's decision in the case.

Related Information: Informative Abstract of the Decision

Summary of Federal District Court's Ruling on FDA's Jurisdiction Over, and Regulation of, Cigarettes and Smokeless Tobacco

May 2, 1997

http://www.fda.gov/opacom/backgrounders/bg97-9.html

On April 25, 1997, Judge William Osteen of the Federal District Court in Greensboro, North Carolina, ruled that FDA has jurisdiction under the Federal Food, Drug, and Cosmetic Act to regulate nicotine-containing cigarettes and smokeless tobacco. The Court held that "tobacco products fit within the FDCA's definitions of drug and device, and that FDA can regulate cigarettes and smokeless tobacco products as drug delivery devices under the combination product and restricted device provisions of the Act.

With respect to the tobacco rule, the Court upheld all restrictions involving youth access and labeling, including two access provisions that went into effect Feb. 28: (1) the prohibition on sales of cigarettes and smokeless tobacco products to children and adolescents under 18, and (2) the requirement that retailers check photo identification of customers who are under 27 years of age.

The Court also upheld additional access and labeling restrictions originally scheduled to go into effect Aug. 28, 1997, including a prohibition on self-service displays and the placement of vending machines where children have access to them. The Court also upheld the ban on distribution of free samples, the sale of so-called kiddie packs of less than 20 cigarettes, and the sale of individual cigarettes. However, the Court delayed implementation of the provisions that have not yet gone into effect pending further action by the Court.

The Court invalidated on statutory grounds FDA's restrictions on the advertising and promotion of cigarettes and smokeless tobacco. Judge Osteen found that the statutory provision relied on by FDA, section 520(e) of the Act (21 U.S.C. 360j(e)), does not provide FDA with authority to regulate the advertising and promotion of tobacco products. Specifically, the Court found that the authority in that section to set "such other conditions" on the sale, distribution, or use of a restricted device does not encompass advertising restrictions. Because Judge Osteen based his ruling on the advertising provisions on purely statutory grounds, he declined to consider the First Amendment challenge to those parts of the rule. The government is appealing the advertising portion of the ruling.

(accessed January 26, 1998)

Related Information: Sample Descriptive Abstract

"Bonanza Creek LTER [Long Term Ecological Research] 1997 Annual Progress Report" http://www.lter.alaska.edu/pubs/1997pr.html

We continue to document all major climatic variables in the uplands and floodplains at Bonanza Creek. In addition, we have documented the successional changes in microclimate in 9 successional upland and floodplain stands at Bonanza Creek (BNZ) and in four elevational locations at Caribou-Poker Creek (CPCRW). A sun photometer is operated cooperatively with NASA to estimate high-latitude atmospheric extinction coefficients for remote-sensing images. Electronic data are collected monthly and loaded into a database which produces monthly summaries. The data are checked for errors, documented, and placed on-line on the BNZ Web page. Climate data for the entire state have been summarized for the period of station records and krieged to produce maps of climate zones for Alaska based on growing-season and annual temperature and precipitation.

Related Information: Sample Informative Abstract based on Experimental Work

Palmquist, M., & Young, R. (1992). The Notion of Giftedness and Student Expectations About Writing. Written Communication, 9(1), 137-168.

Research reported by Daly, Miller, and their colleagues suggests that writing apprehension is related to a number of factors we do not yet fully understand. This study suggests that included among those factors should be the belief that writing ability is a gift. Giftedness, as it is referred to in the study, is roughly equivalent to the Romantic notion of original genius. Results from a survey of 247 postsecondary students enrolled in introductory writing courses at two institutions indicate that higher levels of belief in giftedness are correlated with higher levels of writing apprehension, lower self-assessments of writing ability, lower levels of confidence in achieving proficiency in certain writing activities and genres, and lower self-assessments of prior experience with writing instructors. Significant differences in levels of belief in giftedness were also found among students who differed in their perceptions of the most important purpose for writing, with students who identified "to express your own feelings about something" as the most important purpose for writing having the highest mean level of belief in giftedness. Although the validity of the notion that writing ability is a special gift is not directly addressed, the results suggest that belief in giftedness may have deleterious effects on student writers.

Related Information: Sample Informative Abstract based on Non-experimental Work

Environmental Impact Statement. Federal Register: December 11, 1997 (Volume 62, Number 238). "Endangered and Threatened Wildlife and Plants; Proposed Revision of Special Regulations for the Gray Wolf." Fish and Wildlife Service, Department of the Interior.

http://www.epa.gov/fedrgstr/EPA-SPECIES/1997/December/Day-11/e32440.htm

On November 22, 1994, the U.S. Fish and Wildlife Service published special rules to establish nonessential experimental populations of gray wolves (Canis lupus) in Yellowstone National Park and central Idaho. The nonessential experimental population areas include all of Wyoming, most of Idaho, and much of central and southern Montana. A close reading of the special regulations indicates that, unintentionally, the language reads as though wolf control measures apply only outside of the experimental population area. This proposed revision is intended to amend language in the special regulations so that it clearly applies within the Yellowstone nonessential experimental population area and the central Idaho nonessential experimental population area. This proposed change will not affect any of the assumptions and earlier analysis made in the environmental impact statement or other portions of the special rules. (accessed January 26, 1998)

Related Information: Table of Contents of the Argument

Court Brief (edited Table of Contents) Filed Dec. 2, 1996, by the Department of Justice in defense of FDA's determination of jurisdiction over cigarettes and smokeless tobacco products and its regulations restricting those products to protect children and adolescents. http://www.usdoj.gov/civil/cases/tocnts.htm Statement of the matter before the court; statement of material facts

• The health effects of cigarettes and smokeless tobacco
• The evidence that nicotine in cigarettes and smokeless tobacco "affect[s] the structure or any function of the body"
• The evidence that the pharmacological effects of nicotine in cigarettes and smokeless tobacco are "intended"
• The evidence that cigarettes and smokeless tobacco are "combination products"
• Cigarettes and smokeless tobacco as combination products
• The regulatory goal
• Youth access restrictions
• Advertising and promotion restrictions

Questions Presented

Congress has not precluded FDA from regulating cigarettes and smokeless tobacco under the FDCA.

• Standard of review: Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc.
• Chevron, step one
• Chevron, step two: FDA's application of the FDCA to cigarettes and smokeless tobacco is "based on a permissible construction of the statute"
• No statute, or combination of statutes, can override the FDCA in the absence of express preclusion or other clearly expressed Congressional intent
• Federal Cigarette Labeling and Advertising Act
• Comprehensive Smokeless Tobacco Health Education Act
• Alcohol, Drug Abuse, and Mental Health Administration Reorganization Act
• The separation of powers doctrine does not prohibit FDA's regulation of tobacco products

Nicotine in cigarettes and smokeless tobacco is a drug, and cigarettes and smokeless tobacco are drug delivery devices under the FDCA.

• Cigarettes and smokeless tobacco "affect the structure or any function of the body"
• Nicotine's effects are intended by the manufacturers
• Cigarettes and smokeless tobacco are combination drug/device products and may be regulated under the Act's device authorities
• FDA's application of device provisions to cigarettes and smokeless tobacco is reasonable

The restrictions imposed by FDA on advertising and other promotion of cigarettes and smokeless tobacco are fully consistent with the first amendment.

• The Central Hudson standard and the proper First Amendment analysis
• Recent rulings by the Supreme Court in 44 Liquormart, and by the Fourth Circuit in Anheuser-Busch and Penn Advertising
• In applying the Central Hudson test, the Court's decision should be based on the record created by the Agency, and the reasonable determinations made by FDA are not to be disregarded
• The government's interest here is plainly substantial
• FDA has demonstrated that advertising affects tobacco use by minors, to the detriment of the public health, and that the agency's restrictions on advertising of these products should alleviate that problem to a material degree
• The restrictions are designed to preserve the flow of information to lawful consumers
• The availability of non-Speech related regulatory alternatives does not invalidate FDA's regulations
• Each of FDA's individual advertising restrictions is narrowly tailored

Bibliographic Citation or Identification

As more and more databases are stored and accessed electronically, abstracts are more frequently reproduced apart from the entire article or document. In a large corporation or government entity, for instance, an abstract of a progress report might be circulated and stored in a dozen offices or on multiple computers even though the report itself is filed in only one location. Clear identification is crucial so that readers who want to review the entire text can locate it from the information given with the abstract.

Depending on where your writing is printed and stored, you'll need to include different kinds of identifying information with your abstract:

Bibliographic Citation

If your writing will be printed and disseminated as a book, part of a book, or an article in a journal or magazine, give a full bibliographic citation that includes all the publication information so that readers can find print copies of the article (even if your abstract will appear in unrelated electronic databases). For example, an abstract for a journal article begins with this citation:

Harris, L.D., & Wambeam, C.A. (1996). The Internet-Based Composition Classroom: A Study in Pedagogy. Computers and Composition , 13(3), 353-372.

Organizational Identification

If your abstract is part of a corporate or government document that will not be printed or disseminated outside the organization, you need only include your name, the title of the document, its completion date, a project name (if you produced the document as part of the work on a larger project), and an authorization or organizational number (if there is one).

If your abstract will be circulated outside your organization (for instance, if you work for a consulting company that writes reports for other companies), add to the information above: your company or organization name, the name of the organization that commissioned the document, a contract number (if there is one), a security classification (as appropriate for government documents), and key words to help in cataloguing your abstract.

Internet Citation

If you're "publishing" your own work on the World Wide Web or if your writing will appear on the Internet as part of a full-text electronic database, you can save readers time by citing the Internet address for the full text. Typically, writers note both print publication information and the URL (universal resource locator)--the http or www address--with the abstract.

For example, one of the abstracts cited in this module has this citation that includes both bibliographic information and the Internet address:

Environmental Impact Statement. "Endangered and Threatened Wildlife and Plants; Proposed Revision of Special Regulations for the Gray Wolf." Federal Register: December 11, 1997 (Volume 62, Number 238). Fish and Wildlife Service, Department of the Interior. http://www.epa.gov/fedrgstr/EPA-SPECIES/1997/December/Day-11/e32440.htm

Processes for Writing Abstracts

Unless you work for an abstracting service, you'll usually write abstracts of your own finished work. This section explores some strategies for drafting your abstract.

Cut and Paste Method

Beginning with reading may seem odd since you wrote the paper, but it can frequently be the fastest way to write an abstract because it allows you to "lift" as much of the abstract from your original paper as possible.

• As you read through your own paper, highlight or copy sentences which summarize the entire paper or individual sections or sub-points of your main argument.
• Write (or copy) a sentence that summarizes the main point.
• Add sentences that summarize sections (or write new sentences for sections that lack a concise summary sentence).
• If you're writing a descriptive abstract, you're ready to begin revising.
• If you're writing an informative abstract, look through your paper for details, particularly of key findings or major supporting arguments and major conclusions. Paste these into your abstract and proceed to editing for consistency and length--frequently in the original "cuts" you will still have more detail than is necessary in an abstract.

Outlining Method

Frequently, the best place to start writing an abstract is to first make an outline of the paper to serve as a rough draft of your abstract. The most efficient way to do this is to write what Kenneth Bruffee calls a descriptive or "backwards" outline.

Backwards Outline Instructions

• Read through each paragraph of your paper and write one phrase or sentence that answers the question "what does this paragraph do?"
• Take your list of descriptions for each paragraph and look for connections: i.e., do these 3 or 5 paragraphs do something similar? What is it?
• When you've reduced your outline to 4 or 5 accurate generalizations, you most likely have a descriptive abstract.
• If you're writing an informative abstract, fill in key details about your content.

Detailed Backwards Outline

Because informative abstracts need more detail, the regular backwards outline may not be as useful a strategy for this type of abstract. Instead, do a backwards outline on the left-hand side of a piece of paper. Then, on the right-hand side, answer the question "what does this paragraph say?" for each paragraph in the paper. Then complete the steps below:

• Take your first column and generalize down to 4-5 sentences about what the paper does.
• Use these sentences as topic sentences for the paragraphs in your abstract.
• Now, go to your second column and choose appropriate content for each section you outlined in #2. In other words, use the right-hand column to fill in details about what your paper says on each point outlined in #2.

Key Issues in Preparing Abstracts

Concise, accurate statement of the main idea.

Abstracts begin with a one-sentence summary of the main point of your paper and often introduce the problem the paper explores. Especially for papers based on research, the first sentence (or two) of the abstract announces the subject and scope of the research as well as the problem and your thesis. That's quite a bit of information to condense into a sentence or two, and so the concise statement of the main idea often takes careful revision.

Condensing Information for Non-research Papers

Most non-research papers can be summed up in a nutshell statement—a single sentence that boils down a paper to its essential main point and doesn't aim to capture details, supporting arguments, or types of proof.

One-sentence Summaries for Different Types of Papers Each of these non-research papers summarizes its main point based on its overall purpose:

This paper argues that the "saving democracy" rhetoric surrounding the Gulf War was merely a mask for the U.S.'s interest in keeping oil prices down. (From a political science paper whose purpose was to construct an argument .)

Ethnography and ethnology are the preferred research methods of many anthropologists. (From an anthropology paper whose purpose was to inform others about a research methodology.)

Condensing Information for Research Papers

In addition to stating the main point of the paper, research-based papers often need to set up the context and scope of the research as well. Setting the context includes stating the subject of your work as well as the problem that prompted your research. You might also refer to major researchers who have already done work on your topic as a way of setting the context. Remember, too, that your abstract must always include the main point of your paper, so don't neglect that focus as you work on stating the problem and context. Click on the following links to view examples of condensed statements in research papers:

Related Information: Example 1

In this example, note that the writer uses the names of key researchers to set the context and then focuses on what researchers don't yet know. After setting up the problem he's addressing in the research, the writer then announces the scope and focus of the paper in the second sentence:

Research reported by Daly, Miller, and their colleagues suggests that writing apprehension is related to a number of factors we do not yet fully understand. This study suggests that included among those factors should be the belief that writing ability is a gift. . . .

Related Information: Example 2

In this example, the writer announces the subject and scope of the research although he doesn't set context or suggest the problem that prompted the research. Depending on your ultimate goals for the abstract, you may be more successful with this approach that states the main point of your research paper even without setting context:

This report examines the changes in photosynthesis with an energy-producing carnivorous plants, specifically the Venus Fly Trap. (From a botany research report which involved original lab research.)

Beware of Focusing too Narrowly

No one who has ever written a concise restatement of a complex point will claim that the work was easy or straightforward. Usually, a writer needs to work back and forth between revising the restatement and re-reading the paper to be sure the main idea is stated accurately and clearly. Having worked so hard on that point, though, don't assume that you don't need to revise other parts of your abstract. In this example, the writer restates only the main point and dismisses key information from the 15-page document that should be included in the abstract.

Sample Abstract with Overly Narrow Focus

Community Right-to-Know Notice. Federal Register : January 23, 1998 (Volume 63, Number 15). "Phosphoric Acid; Toxic Chemical Release Reporting." Environmental Protection Agency (EPA).

http://www.epa.gov/fedrgstr/EPA-WASTE/1998/January/Day-23/f1644.htm

EPA is denying a petition to delete phosphoric acid from the reporting requirements under section 313 of the Emergency Planning and Community Right-to-Know Act of 1986 (EPCRA) and section 6607 of the Pollution Prevention Act of 1990 (PPA). This action is based on EPA's conclusion that phosphoric acid does not meet the deletion criteria of EPCRA section 313(d)(3). Specifically, EPA is denying this petition because EPA's review of the petition and available information resulted in the conclusion that phosphoric acid meets the listing criterion in EPCRA section 313(d)(2)(C) in that the phosphates that result from the neutralization of phosphoric acid may cause algal blooms. Algal blooms result in deoxygenation of the water and other effects that may ultimately lead to a number of serious adverse effects on ecosystems, including fish kills and changes in the composition of animal and plant life.

Test Your Ability to Judge Conciseness

The biggest problem writers run into when beginning an abstract is providing enough accurate information to convey an article's main idea without providing more detail than is needed. To test your ability to judge conciseness, read the detailed summary below, and then judge sample restatements of the main idea.

A Detailed Summary

A summary of: Jaime O'Neill, No Allusions in the Classroom, Newsweek , September 23, 1985.

Author Jaime O'Neill's article, "No Allusions in the Classroom," emphasizes the communication problem between teachers and students due to the students' lack of basic knowledge. The author supports this assertion by using a combination of personal experience, evidence obtained from recent polls, other professors' opinions, and the results of an experiment he conducted in his own classroom. The experiment O'Neill conducted was an ungraded eighty-six question "general knowledge" test issued to students on the first day of classes. On this test, "most students answered incorrectly far more often than they answered correctly." Incorrect answers included fallacies such as: "Darwin invented gravity" and "Leningrad was in Jamaica." Compounding the problem, students don't ask questions. This means that their teachers assume they know things that they do not. O'Neill shows the scope of this problem by showing that, according to their teachers, this seems to be a typical problem across the United States. O'Neill feels that common knowledge in a society is essential to communicate. Without this common knowledge, learning is made much more difficult because teacher and student do not have a common body of knowledge from which to draw. The author shows the deterioration of common knowledge through poll results, personal experience, other teachers' opinions, and his own experiment's results.

Related Information: Restatement Test Answers

• Yes - This sentence is inaccurate. While O'Neill was frustrated that his students didn't understand his allusions, the reason he was frustrated is because this lack of knowledge led to a communication problem. The problem of communication in the classroom is O'Neill's main point. The students' misunderstanding of allusions only illustrates this point.
• style="color:#0000ff;"> No - While O'Neill was frustrated that his students didn't understand his allusions, the reason he was frustrated is because this lack of knowledge led to a communication problem. The problem of communication in the classroom is O'Neill's main point. The students' misunderstanding of allusions only illustrates this point.
• Yes - While part of what O'Neill does is expose his students lack of knowledge, he is trying to demonstrate how that lack of knowledge leads to miscommunication. This sentence ignores the communication aspect of his main point.
• style="color:#0000ff;"> No - While part of what O'Neill does is expose his students lack of knowledge, he is trying to demonstrate how that lack of knowledge leads to miscommunication. This sentence ignores the communication aspect of his main point.
• style="color:#ff0000;"> Yes - While this statement is the most accurate so far (because it includes O'Neill's main point that a lack of common knowledge leads to miscommunication in the education process), it fails to show the connection between the two parts of the main point. Nor is it concise, because it separates those two parts of the main point with irrelevant information.
• No - While this statement is the most accurate so far (because it includes O'Neill's main point that a lack of common knowledge leads to miscommunication in the education process), it fails to show the connection between the two parts of the main point. Nor is it concise, because it separates those two parts of the main point with irrelevant information.
• style="color:#ff0000;"> Yes - Though this is the best we've seen so far, it could still be improved slightly by combining the two sentences to make it more concise.
• style="color:#0000ff;"> No - Though this is the best we've seen so far, it could still be improved slightly by combining the two sentences to make it more concise.
• style="color:#ff0000;"> Yes - This is an accurate and concise portrayal of O'Neill's main idea.
• style="color:#0000ff;"> No - This is an accurate and concise portrayal of O'Neill's main idea.

Related Information: Sample Restatements of Main Idea

To test your ability to find a balance between insufficient/inaccurate information and too much information, judge these sample restatements of the main idea of Jaime O'Neill's article.

Organization of Subpoints

After a summary of the main topic/problem/point of your paper or report, the abstract provides some detail on how you reached this point. The information provided in the abstract should follow the organization of the paper/report itself, almost like providing an outline for the reader in text form.

Related Information: Abstracts of Papers With Sub-Headings

When abstracting a paper that has headings and sub-headings, use those to help you identify key parts of the paper for your abstract. The following sample abstract, based on a research paper, uses the introduction, subjects, methods, results, and discussion headings from the original paper.

Note : The numbers in this abstract are for illustration purposes only. Number 1 designates a concise statement of the main point and "problem" prompting the research. Number 2 designates a summary of the selection of research subjects. Numbers 3 and 4 correspond to summaries of research methods and results, respectively, and Number 5 designates a summary of conclusions.

(1) "Students in networked classrooms" examines the question of whether students in a computer classroom are more likely to engage in peer review than students in traditional classrooms. (2) To test this question, two classes in each environment were studied. (3) An observer participated in all four classes for the duration of a semester, noting the nature of the interaction between students. Further, the observer interviewed both students and teachers about the nature of peer interaction and review. (4) Based on this sample, the study finds that students in computer classrooms are more likely, by a ratio of 2:1, to engage in peer review. (5) As a result of this finding, the paper concludes that for this one variable, computer classrooms are a more effective environment in which to teach writing.

Related Information: Abstracts of Papers Without Headings

When abstracting a paper that doesn't have headings and sub-headings, you must depend on your sense of major "chunks" in the text. As you'll see in the following example, this writer followed his concise statement of the main point with two sentences that focus on the two main arguments presented in the paper.

Note: The numbers in this abstract are for illustration purposes only. Number 1 designates a concise statement of the main point. Number 2 designates a summary statement of the first major argument and its support (five pages in the original article). Number 3 corresponds to a summary of the second major argument (two pages in the original), and Number 4 corresponds to the second argument's support (two pages).

(1) This paper argues that the "saving democracy" rhetoric surrounding the Gulf War was merely a mask for the U.S.'s interest in keeping oil prices down. (2) Such an argument is made by first describing the ways in which OPEC controlled oil prices by limiting sales, pointing specifically to how Kuwait was producing more oil than allowed by current OPEC agreements. (3) Second, the paper examines why the U.S. was invested in keeping good relations with the only two OPEC nations--Kuwait and Saudi Arabia--which frequently made trade agreements that benefited the U.S. (4) Finally, the paper does a close reading of the newspaper coverage of the Gulf War, examining how an early recognition of the monetary incentive changed to a democratic one when Bush ordered trOops to Saudi Arabia.

Use of Details

Details should be used judiciously in abstracts. Determining the amount of detail to provide depends a great deal on what type of abstract you are writing (informative or descriptive), the complexity of the paper, the word limit for the abstract, and the purposes you imagine readers of your abstract have for reading.

Complexity of the Paper

An abstract of a five-page progress report is likely to be shorter than an abstract for a 100-page Master's thesis, mainly because a long paper will include more main ideas, not just details. Keep in mind your readers and their reasons for reading your abstract. Focus your abstract on main ideas and provide only those details that are crucial for readers to understand your main points.

Word Limit for the Abstract

Some publications limit the length of abstracts to no more than 75 words. Others allow abstracts of complex documents to run up to 350 words. Be sure to check the publication's guidelines. If it has a low word limit, concentrate on capturing only main ideas from your paper. Don't try to cut a 200-word abstract down to 125 words by simply cutting connecting words, articles, etc. Even the shortest abstracts need to be readable, not telegraphic.

Readers' Purposes

If you're abstracting a report for technical managers, more detail is probably better. But if you're abstracting for a publication, readers will probably skim the abstract to see if they should read the article. Don't give readers more detail than you imagine they'll need to suit their primary goal in reading your abstract.

The five main purposes for abstracts are discussed elsewhere in this guide.

Revising and Editing

When you work from your own texts, abstracts are usually easy to draft. After all, most writers begin by cutting and pasting from the text itself. But abstracts can be tricky to revise and edit, particularly if you need to reach a low word count. In this section, we offer some advice on strategies for moving from a first draft of an abstract to a polished finished version.

Being Concise

When you cut and paste parts of your paper into your draft abstract, you may find that you initially include words and phrases that clarify the meaning in the paper but that simply add extra words in the abstract. Read your drafts carefully to cut unnecessary words. Note that the italicized words in the example can be cut without any loss of meaning in the abstract.

Palmquist, M. (1995). "Students in Networked Classrooms." Computers and Composition, 10 (4), 25-57.

"Students in networked classrooms" examines the question of whether students in a computer classroom are more likely to engage in peer review than students in traditional classrooms. To test this question , two classes in each environment were studied. An observer participated in all four classes for the duration of a semester, noting the nature of the interaction between students. Further, the observer interviewed both students and teachers about the nature of peer interaction and review. Based on this sample , the study finds that students in computer classrooms are more likely, by a ratio of 2:1, to engage in peer review. As a result of this finding , the paper concludes that, for this one variable, computer classrooms are a more effective environment in which to teach writing.

Smoothing out Connections

After you revise for conciseness, you will also want to be sure that each sentence in your abstract leads smoothly into the next. Sometimes you need to add or change transitional words and phrases. Sometimes you need to repeat key words. And sometimes, you need to combine sentences so that the connections between ideas are logically clear.

In our example, we combine what were sentences 2 and 3 and the last two sentences.

This paper examines whether students in a computer classroom are more likely to engage in peer review than students in a traditional classroom. Two classes in each environment were observed, with the participant-observer noting interactions between students. Further, the observer interviewed both students and teachers about peer interaction and review. The study finds that students in computer classrooms are twice as likely to engage in peer review and concludes that, for this one variable, computer classrooms are a more effective environment in which to teach writing.

Avoiding Telegraphic Abstracts

A highly condensed style can save money when you send a telegram but can make abstracts too dense. Don't cut articles ( a, an, the ) or connecting words that show relationships among ideas. Do repeat key words that show the content of your paper. Abstracts may be short, but they are meant to be readable.

Polishing Style

A reader looks at a summary for the sole purpose of getting a quick glimpse of the article. As a result, she doesn't want to waste time with a lot of phrases and words that do not further the meaning, nor is she interested in the summary writer's opinion. Accounting for audience needs, there are three generalizable principles about the style of summaries:

Use of "I"

Although use of "I" or "we" is acceptable in some disciplines, many frown on its use in abstracts. Read several abstracts in the publication you're submitting to or the databases you expect to include your abstract. When in doubt, do not use "I." Instead, use the following strategies:

Substitute for "I" Most abstracts make the paper/report/study the focus of the abstract and the grammatical subject of sentences in the abstract. Try these sentence openers:

• This paper explores. . .
• This study suggests. . .
• The report investigates. . . .

Passive Voice In combination with substitutes for "I," passive voice helps writers focus on the paper/report/study. Instead of, "I propose that ethnography is a better research method than case study" (active voice), the abstract might use: "Ethnography is proposed as a better research method than case study." (passive voice) Be sure to combine substitutes for "I" with passive voice to avoid overusing the passive.

Use of Quotes

When using your own sentences, you don't need to put them in quotation marks. For example, if your methods section begins with "Three methods were used to investigate this question: case study, surveys, and observational research," feel free to repeat the sentence in its entirety in the abstract. Remember, however, the following points:

• Revise the sentence so it makes sense in the abstract (i.e., if you have not summarized "this question" in the abstract, omit substitute for that phrase).
• Do not "lift" sentences which are not your own (i.e., quotes from other people's work).

Use of Literary Present Tense

Abstracts use the present tense because we assume texts speak to the present even if their authors are dead and/or wrote the words in the past. As a result, write about the text and/or author as if they were composing the words at the moment. For example:

• Hemingway describes Paris as......
• The Declaration of Independence states that all men are created equal.

Caution: This rule varies from discipline to discipline.

Abstracts in Specific Disciplines

Abstracts have common elements and uses, but read enough abstracts in your field to be aware of their specific details or differences. Choose from the examples to see additional sample abstracts. The abstract from Civil Engineering includes instructor comments.

Civil Engineering

MASK Engineering has designed a performing arts center for the CSU campus in order to provide a complex that will better serve the campus and the community. This facility will not only improve the performing arts programs on campus but will encourage students and community members to attend more cultural events in Fort Collins. The capacity of the new facility will exceed that of existing structures on campus, and the quality of sound and aesthetics will be improved. Some of the features included are a large performing hall, a coffee shop, a banquet hall, and a recording studio. The total area of the complex is 56,500 square feet split into three levels.

Instructor Comments

This abstract summarizes the accomplishments of the project and what it will do. It also summarizes some of the actual design and indicates that it's going to include a performing hall, coffee shop, banquet hall, and recording studio.

The writing, however, could be a little tighter in my opinion. The first sentence looks like it's around 20 words long. First of all, the expression "will better service the campus and the community" doesn't mean anything. What does "better serve" mean? A better choice might be, "MASK Engineering has designed a new Performing Arts Center that will meet the needs of the theater community," or something more specific.

The second sentence is typical. It gives the particular vehicle for doing the programs. However, it implies that the facility improves programs, and I'm not sure that's quite the right subject for this sentence. Furthermore, there's no point to the word "but" here. There's no contrast here, so this is a grammatical problem. This kind of problem can be avoided through careful reading, asking what each sentence accomplishes.

The abstract gets stronger after this. "The capacity of the new facility will exceed that" is very specific. "The quality, sound and aesthetics will be improved. Some of the features included are this..." The writers are very good at being descriptive. I think engineering students are more comfortable with the descriptive aspect of their material than with the lead-in.

LeCourt, D. (1996. Composition's Theoretical Irony: WAC as Uncritical Pedagogy. Journal of Advanced Composition, 16 (3), 389-406.

This paper argues that writing across the curriculum has failed to consider how its practices and theories serve to inscribe students within normalized discourses. As scholars such as Susan McLeod, Anne Herrington and Charles Moran begin to re-think the way writing-across-the-curriculum programs have situated themselves within composition theory, an intriguing disparity has presented itself between writing-to-learn and learning-to-write. As McLeod points out, these two approaches to WAC, which she designates the "cognitive" and the "rhetorical," respectively, exist in most programs simultaneously despite their radically different epistemological assumptions. This paper suggests, however, that despite the two approaches' seeming epistemological differences, they work toward a similar goal: the accommodation or inscription of (student) subjects into the various disciplinary strands of academic discourse. From a poststructural perspective, the goals of both these models function as a coherent technology of subject production. Writing to learn exercises provide a discursive space in which students learn to write themselves as subjects of the discourse, using the writing space to "practice" an integration of self with a disciplinary subjectivity. The rhetorical model reinforces such an integration even more strongly, providing explicit instruction in how the discursive subject must write herself in order to produce "effective" prose which mirrors the texts of other "speaking" subjects of the discourse. In sum, both approaches to WAC are subject to the same description and critique of how academic discourse seeks to inscribe students as subjects that has been forged against composition instruction in English departments (e.g., Schilb, Clifford, Faigley). Ironically, in WAC, we have presumed a clear mission for writing instruction that is not nearly so evident in our own approach to advanced literacy. The paper concludes, then, by offering yet a third model of WAC, one which suggests that students, as well as their instructors, engage in the investigative process of discovering how discursive conventions relate to their discipline's epistemology and consider how that connection limits what can be said or thought within that discourse.

Neurobiology

High Performance Computing Applications in Neurobiological Research; Muriel D. Ross, NASA Ames Research Center, Moffett Field, CA 94035; Kevin Montgomery, Sterling Software, Palo Alto, CA 94303; David G. Doshay, Sterling Software, Palo Alto, CA 94303; Thomas C. Chimento, Sterling Software, Palo Alto, CA 94303; Bruce R. Parnas, National Research Council Research Associate, Biocomputation Center, NASA Ames Research Center, Moffett Field, CA 94035

The human nervous system is a massively parallel processor of information. The vast numbers of neurons, synapses and circuits is daunting to those seeking to understand the neural basis of consciousness and intellect. Pervading obstacles are lack of knowledge of the detailed, three-dimensional (3-D) organization of even a simple neural system and the paucity of large scale, biologically relevant computer simulations. We use high performance graphics workstations and supercomputers to study the 3-D organization of gravity sensors as a prototypic architecture foreshadowing more complex systems. Scaled-down simulations run on a Silicon Graphics workstation and scaled-up, three-dimensional versions run on the Cray Y-MP and CM5 supercomputers.

To assist this research, we developed generalized computer-based methods for semiautomated, 3-D reconstruction of this tissue from transmission electron microscope (TEM) serial sections and for simulations of the reconstructed neurons and circuits. Sections are digitized directly from the TEM. Contours of objects are traced on the computer screen. Mosaicking images into sections, registration and visualization are automated. The same grids generated to connect contours for viewing objects provide tesselated surfaces for 1-D, 2-D and 3-D simulations of neuronal functioning. Finite element analysis of prism or segment volumes and color coding are used to track current spread after synapse activation. The biologically accurate simulation is reducible to a symbolic model that mimics the flow of information processing. Discharge patterns are displayed as spike trains. The symbolic model can be converted to an electronic circuit for potential implementation as a chip. The reconstructions can also be rendered in visual, sonic and tactile virtual media.

Using these methods, we demonstrated that gravity sensors are organized for parallel distributed processing of information. They have non-modular receptive fields that are organized into overlapping, dynamic cell assemblies. These provide a basis for functional degeneracy and graceful degradation. The sensors have two intrinsic microcircuits that are prototypic of more advanced systems. These microcircuits are highly channeled (type I cell to a nerve terminal called a calyx) and distributed modifying (type II cells and feedforward/feedback neural lOops). A circuit of extrinsic origin likely biases the intrinsic circuits. We use simulation methods to study the effects of intrinsic feedback-feedforward lOops and of extrinsically driven biases on discharge patterns. These and similar investigations into the functioning of huge assemblies of neurons require supercomputer capabilities and pave the way for studies of human brain functioning as a grand challenge in supercomputer applications.

http://biocomp.arc.nasa.gov/papers/hpc_abstract.94.html

(accessed February 3, 1998)

Department of the Interior - U.S. Geological Survey

Inventory of Landslides Triggered by the 1994 Northridge, California Earthquake Edwin L. Harp and Randall W. Jibson Open-File Report 95-213 USGS Denver, CO 80225 1995

The 17 January 1994 Northridge, California, earthquake (M=6.7) triggered more than 11,000 landslides over an area of about 10,000 km. Most of the landslides were concentrated in a 1,000-km area that includes the Santa Susana Mountains and the mountains north of the Santa Clara River valley. We mapped landslides triggered by the earthquake in the field and from 1:60,000-scale aerial photography provided by the U.S. Air Force and taken the morning of the earthquake; these were subsequently digitized and plotted in a GIS-based format, as shown on the accompanying maps (which also are accessible via Internet). Most of the triggered landslides were shallow (1-5 m), highly disrupted falls and slides in weakly cemented Tertiary to Pleistocene clastic sediment. Average volumes of these types of landslides were less than 1,000 m, but many had volumes exceeding 100,000 m. Many of the larger disrupted slides traveled more than 50 m, and a few moved as far as 200 m from the bases of steep parent slopes. Deeper ( >5 m) rotational slumps and block slides numbered in the hundreds, a few of which exceeded 100,000 m in volume. The largest triggered landslide was a block slide having a volume of 8X10E06 m. Triggered landslides damaged or destroyed dozens of homes, blocked roads, and damaged oil-field infrastructure. Analysis of landslide distribution with respect to variations in (1) landslide susceptibility and (2) strong shaking recorded by hundreds of instruments will form the basis of a seismic landslide hazard analysis of the Los Angeles area.

http://gldage.cr.usgs.gov/html_ files/ofr95-213/ABSTRAC2.html

LeCourt, Donna,  Kate Kiefer, Luann Barnes, Mike Palmquist, & Tom Siller. (2004). Abstracts. Writing@CSU . Colorado State University. https://writing.colostate.edu/guides/guide.cfm?guideid=59

Search form

How to write an abstract.

Write with Might #30: How to Write an Abstract

This week's writing tip focuses on  how to write an abstract , according to APA style. I understand that first-yearMSWs are finalizing their research projects this quarter and may appreciate a quick conversation on this aspect of research manuscripts. For those who are not currently working in research, having the ability to write an abstract is a valuable skill as we know that involvement in social work research has the power to change systems in the direction of social justice.  The following information is adapted from the Purdue OWL website, with citation following.

Why an Abstract?

An abstract gives your reader a brief summary of the contents of your research manuscript. The process of writing it can actually be helpful to the writer as well! Consider writing your abstract before you have finished your first draft as the process can give you feedback on the clarity of your arguments and the organization of your thoughts.

Abstracts can also alert your reader to keywords from your work. Including keywords on your abstract page allows readers to locate your paper in databases with ease.

What Does it Include?

An abstract includes a succinct summary of the main ideas in your work. When writing the abstract, focus on including: the research topic, research questions, participants, methods, results, data analysis and conclusions. Abstracts may also speak to the possible implications of your research and the direction you recommend for future research related to your work.

APA Formatting Details?

-The abstract lives on its own page that includes the header found throughout your paper.

-It is a single, double-spaced paragraph between 150-250 words.

-Write the word “Abstract” on the the first line of the abstract page. The word "Abstract" should be centered and in regular type. Do NOT use boldface, formatting extras, italics, underlining, or quotation marks, etc.

-To include your paper's keywords, write "keywords" indented and in italics with a colon following it. Then type the keywords in regular text.

Paiz, J., Angeli, E., Wagner, J., Lawrick, E., Moore, K., Anderson, M., Soderlund, L., Brizee, A., Keck, R. (March 1, 2013). General Format. Retrieved from: http://owl.english.purdue.edu/owl/resource/560/01/

Sample abstracts

Reprocessing used nuclear fuel (UNF) is crucial to the completion of a closed fuel cycle and would reduce the volume of waste produced during nuclear power production. Pyroprocessing is a promising reprocessing technique as it offers pure forms of product recovery. A limiting issue with pyroprocessing, however, is the inability to monitor concentrations of chemical species inside the electrorefiner. As with many nuclear processes, safe guards and monitoring become increasingly important; therefore, development of real - time monitoring techniques for various chemical species may allow for commercialization of this recycling process [1 - 5]. The focus of the proposed research is to develop accurate diffusion coefficients for Yttrium, a fission product found in UNF, in molten salt conditions through Cyclic Voltammetry (CV). Quantification of the diffusion coefficient will allow current measurements from inside the melt to be directly related to species concentration. With the diffusion coefficients, in - situ CV would then facilitate real - time monitoring of chemical concentrations.

This project aims to analyze the social and cultural effects of the Iranian Revolution through primary source material and interviews with those directly affected by the revolution. Iran’s political seclusion and its animosity toward the West has limited the voices and perspectives available to an American audience. Moreover, the attitude of the West towards Iran since the revolution has been myopic and often marred by political perspectives. The objective of this project will be to bring those voices and stories to light, putting a greater focus on the experiences of individuals who lived through the Revolution. These stories will be presented in a digital medium (film and web) in order to bring these voices and perspectives to an American audience.

https://www.youtube.com/watch?v=XqeIIjRB3EU

Special Considerations for UURAF and Mid-SURE Abstracts

Every UURAF and Mid-SURE submission requires an abstract, which are published in the online program book. Visitors to UURAF and Mid-SURE will review your abstract when deciding which presentations to attend, so it is important that the abstract follows the appropriate format and succinctly describes your research. Here are some additional points to consider when writing your UURAF or Mid-SURE abstract:

• Abstracts are limited to 250 words in length. Any over the word limit will be truncated in the program booklet.
• Submit abstracts as one paragraph with no breaks and no footnotes.
• Draft versions are accepted when registering for UURAF and Mid-SURE. You will have an opportunity to provide a final version for the program booklet.
• Many visitors to UURAF and Mid-SURE will be unfamiliar with your discipline and its vocabulary. Take this into consideration when writing your abstract.
• Do not include symbols or a course number (e.g., EAD 315) in the title.
• Symbols (e.g., Δ, σ, β, °, ≤, ±) and italics may not transfer correctly into the UURAF or Mid-SURE database. If your abstract contains symbols or italics, please upload a PDF of your abstract during registration or email to [email protected].
• We encourage all students to review their abstracts with their research mentors to make sure that copyright or intellectual property (IP) issues are not a factor.
• Check out our Abstract FAQ to learn more about abstracts and how to prepare them
• Get help from an Undergraduate Research Peer Advisor

• Call us: (517) 884-4384
• Contact Information
• Privacy Statement
• Site Accessibility
• Call MSU: (517) 355-1855
• Visit: msu.edu
• MSU is an affirmative-action, equal-opportunity employer.
• Notice of Nondiscrimination
• Spartans Will.
• © Michigan State University

How to Write an Abstract Step-by-Step: a Guide + Examples

Writing an abstract is one of the skills you need to master to succeed in your studies. An abstract is a summary of an academic text . It contains information about the aims and the outcomes of the research. The primary purpose of an abstract is to help readers understand what a particular paper is about. It serves as a sort of introduction to the paper. The usual length is about 150-300 words.

Our specialists will write a custom essay specially for you!

Types of abstract essays include descriptive, critical, highlight, and informative abstract.

This article by Custom Writing experts will help you write a perfect abstract. Not only we have an example of informative abstract but also the examples of other types too. Keep reading and good luck with your assignment!

• 🚦 Getting Started

Informative Abstract

Descriptive abstract.

• 👣 Step-by-Step Guide
• 🔗 References

🚦 How to Write an Abstract: Getting Started

There are several things to consider before you start writing an abstract.

• It would be best if you had your paper ready. This one should be a no-brainer, but it’s still worth mentioning. If you try to write your abstract first, chances are you’ll have to edit it a lot afterward.
• Make sure you’re aware of all the requirements : writing style, length, and the whole purpose of an abstract. All of these factors will influence the contents of your abstract. Again, it’s better to do everything right from the beginning than to edit your work later.
• Think of the audience . Remember the definition of an abstract? It helps readers understand what your work is about. You need to be aware of who’s going to read it. Are they going to be scientists who’ll use your abstract to decide whether your work is relevant? Or do you need to make your abstract easy to understand for anyone? Answering these kinds of questions will help you determine how your abstract will look.
• Decide on the type of abstract . This decision is an essential one. Therefore, we’ll talk about it in the next part of our guide.

📑 Abstract Structure & Types

There are two main types of abstracts: informative and descriptive. The former is also known as a complete abstract, while the latter contains less information. See more detailed information below.

This type of abstract writing is also known as a complete abstract . And it’s pretty self-explanatory. An informative abstract is a summary of a paper. It describes its purpose, methodology, background, results, and conclusion. It also includes information about the paper’s structure, its key thoughts, and the major topics discussed. How long should an informative abstract be? It usually sticks to around 250+ words. The completeness of the information provided in it makes it possible to use the informative abstract as an independent document. A format similar to informative abstracts is used to write short scientific reports. Apart from examples below, you can use a summary writing tool to generate your own and check out the structure using more materials.

Just in 1 hour! We will write you a plagiarism-free paper in hardly more than 1 hour

Informative Abstract Example

The Internet of Things Provides Emerging Business Opportunities

Purpose: The Internet of Things (IoT) is a new phenomenon, so there is a lack of public and scientific understanding of what IoT is and what commercial opportunities it can offer for large companies and sole entrepreneurs. The article aims to stimulate creativity, thinking, and entrepreneurship in terms of IoT.

Methods: This article consists of three parts. In the first part, IoT is described as a wide socio-technical phenomenon. Second, this article suggests two approaches for establishing new business models using IoT: a disruptive and a sustainable approach. Third, the article concludes with a reflection on the time to which the future of IoT can be possibly predicted.

Scope: The article discusses different elements that comprise IoT in the physical, technological, and socioeconomic environments. Results: This discussion shows the limitations of the new business models approach that was examined in this article and suggests guidelines on the more efficient ways of using this approach.

Conclusions: The attempt to predict the future can prepare companies for various threats and opportunities. The envisioned outcomes and scenarios can help the entrepreneurs make the correct decisions for their businesses’ success.

This type of abstract is also called an indicative abstract, or a limited abstract. Again, the name says it all. This abstract type paints a general description of the paper without going into very in-depth details. In the case of an informative abstract, you can develop an opinion about the paper based on the abstract alone. With a descriptive abstract, though, you’ll still have to read the main work because the abstract will only provide a general idea without all the vital pieces of content. It’s more like a table of contents but written in the form of a paragraph. And it’s usually about 100-200 words long.

Descriptive Abstract Example

Exploring the Boundaries of the Social Sciences

Purpose: The concept of research boundaries has been critical in history, anthropology, sociology, social psychology, political science, and sociology. This article intends to explore this problem and analyze the relational processes hindered by the boundaries.

Methods: It addresses relatable processes in various research institutions and social locations. It also investigates the directions for further development, with a focus on the dependence between symbolic and social boundaries, their cultural mechanisms, hybridity and difference, and group classifications.

Scope: The article analyzes several works on social identity; class, ethnic, and gender inequality; professions and science; and national identities, communities, and territorial boundaries.

👣 Writing an Abstract Step by Step

You’ll need to write an abstract for almost any academic text: a thesis, a research paper, an article, etc. No matter what document you are working on, the abstract should be the last part you’ll write. Let’s learn what main components that any abstract contains and how to write them step by step.

Identify Your Aims

Tell your readers why your work matters and why it is important. Don’t go into details here. Concentrate on the crucial points. Note that this part should be written in the present or past simple tense, not the future, as your research is already done. The questions below can help you formulate your aims.

• Why did I decide to study this particular topic?
• What theoretical or practical problem does my research respond to?
• What is the social context of my work?
• Why are my key findings important?

Explain Your Methods

The next part of your abstract is to contain a short and straightforward description of your research. Explain what you did in one or two sentences. Do that using the past simple tense.

• Describe your research process. Mention the approach you decided to go with and all the data that was at your disposal.
• Give a short overview of the most important sources used for your paper.
• Mention the evidence that supports your claims, so the readers know there’s a foundation to what you’re saying.

Share Your Results

This is where the main difference between the two types of abstracts comes into play.

Receive a plagiarism-free paper tailored to your instructions. Cut 15% off your first order!

You see, it’s only informative abstracts that contain this information. So, if you’ve decided to go with the descriptive type, you may skip this step.

If stating the problem can be considered a question, then this part is the answer to that question. Describe all your general findings as well as the goal that you reached through your research. Support your words with arguments and hypotheses.

Write a Conclusion

Not only will this part be a logical finish to your abstract, but it will also make a smooth transition to its closure.

Explain what your findings mean and why they make your paper important. To simplify the task, use an article summary generator and just edit the resulting piece.

While this part is necessary both for informative and descriptive abstracts, it’s only the former that needs to answer the following questions:

Get an originally-written paper according to your instructions!

• What implications does my work have?
• Are my findings specific or general?

✏️ Abstract FAQ

An abstract is a compressed view of the essential elements of a manuscript without added interpretation. It should consist of an introduction or background, purpose, methodology, results, and discussion, or another conclusion. An effective abstract is not a selection of manuscript sentences, but a reworded gist.

A presentation abstract shall provide an overview of the research as briefly as possible. It shall comprise context, objective, methodology, and findings. The total word count shall not exceed 250 words. You will present the research orally and visually at the conference, so the purpose is to raise the listeners’ curiosity rather than provide them with a summary of your work.

It should be an easy-to-read 250-word passage following the template:

• The topic and purpose of your research or invention.
• The problem you resolved or the hypothesis you examined.
• The scientific methods you used to implement point 2.
• The achieved results.
• Conclusions on the relevance and importance of your project.

Identify the problem you address and give the reasons that motivated you to conduct the investigation. Mention the gaps that require further research. Describe your methodology. Provide the main results and findings without further explanation. The total number of words in an abstract shall be given in the lab requirements.

• How to Write an APA Abstract
• How to Write an Abstract—Parts of an Abstract and Checklist
• What are Structured Abstracts?
• Writing Report Abstracts
• Formatting Your Paper
• Make a Great First Impression: 6 Tips for Writing a Strong Abstract
• Science Fair Project Abstract Writing
• Writing a Research Abstract
• What Goes in an Abstract?
• Share to Facebook
• Share to Twitter
• Share to LinkedIn
• Share to email

So you have to write a literature review. You find your favorite novel and then start analyzing it. This is how it’s usually done, right? It’s not. You have to learn the elements of literature review and how to deal with them.

Looking for relevant sources of information for your research is already a tiring process, but when you also need to pay attention to their credibility, it becomes almost impossible! However, it’s still a quite critical aspect to pay attention to. Using unreliable sites like Wikipedia, even if it’s a two-page...

Only two words, but you already feel a chill down your spine. A research paper is no joke. It’s a super detailed piece of academic writing where you analyze a chosen issue in-depth. The main aim of such torture is to show how knowledgeable you are and that your opinion...

A research proposal is a text that suggests a topic or research problem, justifies the need to study it, and describes the ways and methods of conducting the study. Scholars usually write proposals to get funding for their research. In their turn, students might have to do that to get...

It’s longer than your Bachelor’s thesis. It’s more stressful. It’s more important. And you have no clue how to write it. We understand that a lot comes with the responsibility of creating a Master’s thesis from scratch. But no need to stress out; you can get all the help you...

Students struggle with academic writing rules. Mastering them takes longer than writing the paper. But when the conventions become a normal state of things, the process becomes faster and easier than before. The less you think about the form, the more time you have to think about the contents.

Few students fully get the meaning and the importance of a research proposal. If you have a good research proposal, it means that you are going to carry out adequate research. A low-quality research proposal may be the reason your research will never start.

A lab report is quite a serious piece of paper that has a massive value in your research. And don’t be deceived by the name as the lab report is not a form you just need to fill in. There is an impressive list of components you need to describe...

A literature review is an update on the status of current research related to the issue in question. Its purpose is to provide the reader with a guide to a particular research topic. And for the writer, a well-written literature review is the best way to show their competence in...

A research proposal is a document that aims to show the significance and value of a particular project. It is common to have to write research proposals to acquire funding for various research projects. But that’s not all. Perhaps the most important function of a research proposal is that it...

Are you tired of getting average grades for your term papers? Well, that’s good that you’re here! It may be true that some excellent writing and time-management skills can help you create an impressive piece, but it’s not necessary. Most of the assignments have similar outlines, whether it’s a term...

Congratulations! A dissertation is your last step before getting your doctoral degree. But, no matter how great the excitement is, the frustration and panic might be overwhelming. And it’s understandable as there is a lot of pressure on you right now. The good news is that there is nothing to...

Great article!

Thanks for sharing this blog with us. I am the regular reader of this website. Whenever I visited here find something meaningful. I will share this blog with my friends.

I have found this guide helpful and will be sharing the link with my students.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Data Descriptor
• Open access
• Published: 02 May 2024

Mapping annual 10-m soybean cropland with spatiotemporal sample migration

• Hongchi Zhang   ORCID: orcid.org/0009-0006-5741-7984 1 , 2 , 3   na1 ,
• Zihang Lou 1 , 2 , 3   na1 ,
• Dailiang Peng 1 , 2 ,
• Bing Zhang   ORCID: orcid.org/0000-0003-0319-7753 1 , 3 ,
• Wang Luo 4 ,
• Jianxi Huang   ORCID: orcid.org/0000-0003-0341-1983 5 ,
• Xiaoyang Zhang   ORCID: orcid.org/0000-0001-8456-0547 6 ,
• Le Yu   ORCID: orcid.org/0000-0003-3115-2042 7 ,
• Fumin Wang 8 ,
• Linsheng Huang 9 ,
• Guohua Liu 10 ,
• Shuang Gao 10 ,
• Jinkang Hu 1 , 2 , 3 ,
• Songlin Yang 1 , 2 , 3 &
• Enhui Cheng 1 , 2 , 3  

Scientific Data volume  11 , Article number:  439 ( 2024 ) Cite this article

215 Accesses

Metrics details

• Environmental sciences

China, as the world’s biggest soybean importer and fourth-largest producer, needs accurate mapping of its planting areas for global food supply stability. The challenge lies in gathering and collating ground survey data for different crops. We proposed a spatiotemporal migration method leveraging vegetation indices’ temporal characteristics. This method uses a feature space of six integrals from the crops’ phenological curves and a concavity-convexity index to distinguish soybean and non-soybean samples in cropland. Using a limited number of actual samples and our method, we extracted features from optical time-series images throughout the soybean growing season. The cloud and rain-affected data were supplemented with SAR data. We then used the random forest algorithm for classification. Consequently, we developed the 10-meter resolution ChinaSoybean10 maps for the ten primary soybean-producing provinces from 2019 to 2022. The map showed an overall accuracy of about 93%, aligning significantly with the statistical yearbook data, confirming its reliability. This research aids soybean growth monitoring, yield estimation, strategy development, resource management, and food scarcity mitigation, and promotes sustainable agriculture.

Similar content being viewed by others

Land conversion to agriculture induces taxonomic homogenization of soil microbial communities globally

Critical transitions in the Amazon forest system

A global meta-analysis on the effects of organic and inorganic fertilization on grasslands and croplands

Background & summary.

Soybeans ( Glycine max ) are extensively grown for their high oil content, abundant protein, and substantial contribution to energy production 1 . Over the last two decades, soybeans have consistently played a vital role in the Chinese diet 1 and have been a crucial source of oil and animal feed. China is the world’s largest consumer of soybeans 2 . China produced 20.28 million tons of soybeans in 2022 while importing an additional 91.08 million tons from countries such as Brazil, the United States, and Argentina 3 . Forecasts indicate that China’s soybean demand will reach around 133 million tons by 2035, increasing the pressure on domestic production and imports 4 , 5 . Utilizing satellite-based earth observation data for national-scale mapping of soybean is a cost-effective method to gather comprehensive information 6 , 7 . This spatial information can effectively reveal soybeans distribution, laying a strong foundation for agricultural management and yield prediction 6 , 7 .

Since the late 1990s, remote sensing imagery has progressively assumed a pivotal role in the identification and monitoring of crops 8 , 9 . Numerous researchers have undertaken nationwide and regional-scale crop mapping utilizing remote sensing data 10 , 11 , 12 , 13 . In the early years, researchers typically utilized single-phase or multi-temporal images as primary data sources for crop remote sensing identification, obtaining one or more images during the critical growing season to facilitate crop identification 14 , 15 , 16 . This method has a small amount of data and computational complexity, but the crop identification features extracted are relatively limited, so the accuracy is relatively low. Time-series data has garnered increased attention in recent years for crop identification due to its capacity to capture crop growth patterns accurately. Several studies have employed time-series data for precise crop identification 17 , 18 , 19 , 20 , 21 . Regarding classification, machine learning algorithms, with their robust self-learning and generalization capabilities, have consistently exhibited exceptional accuracy and stability in classifying crops through remote sensing, rendering them among the most widely utilized techniques 22 , 23 . Despite substantial advancements in data and methods for remote sensing crop classification, accurately distinguishing specific crop types such as soybeans, corn, and wheat from imagery remains a formidable challenging 24 .

While various crops exhibit categorical differences, their shared vegetative characteristics often lead to subtle spectral distinctions. Consequently, it is essential to incorporate vegetation index and specific spectral bands to capture the distinct biophysical attributes of crops, particularly soybeans, which frequently encounter pronounced spectral overlap with certain other crops 8 . Previous studies have highlighted a noteworthy feature of soybeans related to reduced canopy water content during the growing season, distinguishing them from some other crops at comparable phenological stages 25 , 26 . The short-wave infrared (SWIR) bands effectively capture this information 27 , 28 . Additionally, the red-edge bands (Sentinel-2) and vegetation index derived from these bands, such as Red Edge Normalized Difference Vegetation Index (RENDVI) and Red Edge Position Index (REPI), play a crucial role in discriminating soybean from corn, thereby enhancing the classifier’s accuracy in soybean classification 20 , 29 .

Methods for crop recognition relying on spectral or vegetation indexes as input features often depend on specific datasets and ground references. However, obtaining sufficient ground-truth crop data typically constitutes the most demanding, time-consuming, and expensive aspect of crop mapping 30 . Consequently, numerous researchers have focused on studying crop mapping in scenarios with either no samples or limited samples. For instance, researchers have explored the physicochemical characteristics of crops by analysing their spectral and vegetation index profiles. They have developed techniques such as knowledge transfer topologies 31 , multi-temporal Gaussian mixture models 32 , and the integrated Greenness and Water Content Composite Index (GWCCI) 27 for mapping soybeans and corn in space. However, these methods are often more suitable for regions with extensive soybean or corn cultivation and may be sensitive to other types of vegetation or crops in areas with intricate planting patterns. Consequently, some researchers have chosen to employ crop ground-truth samples from preceding years for feature transfer or sample migration 33 , 34 . Despite potential variations in certain crop features across time and space, these characteristics tend to exhibit a consistent level of stability 35 . Supervised learning methods are subsequently employed to conduct crop classification in subsequent years 22 , 33 , 36 . Results obtained through these limited or zero-sample methods may not be optimal but still demonstrate acceptable performance and accuracy.

Developing nationwide crop maps presents a formidable challenge that necessitates the availability of high-quality remote sensing data, abundant ground-truth crop data, and well-designed classification methods 37 . In China, soybean cultivation spans approximately 8% of arable land, with nearly half concentrated in the northeastern region 3 . However, in the northeast, soybeans cultivation merely encompasses 5% of the available arable land 38 . These factors highlight the distinctive nature of soybean cultivation in China, characterized by a small cultivation area, dispersed plots, and considerable annual variations. Consequently, the creation of nationwide soybean distribution maps is a highly intricate undertaking. Majority of the current spatial maps for soybean in China are primarily concentrated in the northeastern region 24 , 25 , 39 , and only one product in 2019 covers the whole country (GLAD maize and soybean map) 13 . One of the primary obstacles in generating high-resolution soybean maps lies in the absence of reliable ground-truth data. While some methods have been devised to classify and map soybeans with minimal or no samples 27 , 31 , as well as perform early-season classification using data using previous years 22 , these approaches possess limitations, particularly in the diverse crop landscape of Huang-Huai-Hai Plain and the Middle-Lower Yangtze Plain. Moreover, intricate planting practices on small farms are affected by various factors, including economic shifts and alterations in land use policies 40 , 41 , thereby leading to annual variations in crop types and rendering crop identification more challenging. Despite earnest efforts to map crops at a 10-meter resolution throughout China, soybean mapping remains restricted, especially in regions like Sichuan, Anhui, and Henan, which boast high levels of soybean production and lie beyond the primary Northeastern cultivation areas.

In response to the challenges posed by the lack of high spatial resolution soybean mapping and the absence of ground truth samples in China, our study aims to create China’s annual soybean map based on limited samples and spatiotemporal migration methods. Initially, based on the growth physical and chemical characteristics of soybeans, we generate samples from limited ground survey samples for the target year and region. We then employed random forest classification, utilizing soybean temporal features derived from time series of vegetation indexes and spectral bands as input. These features encompassed statistical measures, phenological characteristics, and harmonic fitting parameters. In regions with frequent cloud cover or a shorter soybean growing season, SAR data were incorporated to complement features. This involved utilizing statistical features and principal component features of backscatter coefficients and their combinations. Our comprehensive approach enables nationwide mapping of soybean planting areas. We successfully generated spatial maps of soybean cultivation for 10 provinces in China, including Heilongjiang, Jilin, Inner Mongolia, Henan, Sichuan, and others, covering the years from 2019 to 2022 at a 10-meter resolution.

Our soybean cropland mapping process includes four steps, as illustrated in Fig.  1 : data preprocessing, sample generation, classification and validation.

Workflow for mapping soybean planting areas using the sample-generation and pixel-based algorithm. Sentinel-2 SR, sentinel-2 surface reflectance products in Google Earth Engine; Sentinel-1 SAR, a dual-polarization C-band Synthetic Aperture Radar data at 5.405 GHz; S-G filter, Savitzky-Golay filter; ESA, European Space Agency; VIs, Vegetation Indices; RMSE, root-mean-square error; MAE, mean absolute error; R2, R-squared; OA, overall accuracy; PA, producer’s accuracy; UA, user’s accuracy.

This study aimed to map the distribution of the soybean planting areas from 2019 to 2022 in ten provinces, namely Heilongjiang, Inner Mongolia, Anhui, Sichuan, Henan, Jilin, Jiangsu, Shandong, Hubei, and Liaoning (Fig.  2 ). These provinces are recognized as the top soybean-producing regions in China, collectively accounting for more than 80% of soybean production 3 . To effectively map the soybean annual planting area in China, we categorized them into three main regions: (a) Northeast China, encompassing Heilongjiang, Jilin, Liaoning, and the northeastern Inner Mongolia; (b) Huang-Huai-Hai Plain and the Middle-Lower Yangtze Plain, covering Shandong, Henan, Anhui, Jiangsu, and Hubei; and (c) Sichuan Basin.

Location of the major soybean-producing region in ( a ) Northeast China, ( b ) Huang-Huai-Hai Plain and the Middle-Lower Yangtze Plain, and ( c ) Sichuan Basin.

The Sentinel-2 satellite, with a spatial resolution of 10 meters and a revisiting period of 5 days, offers optimal support for the comprehensive and long-term identification of crops. Equipped with the Multispectral Instrument (MSI), Sentinel-2 can effectively even the slightest variations between different crops. For our study area, we acquired all available Sentinel-2A/B (S2) Level-2A surface reflectance (SR) data from 2019 to 2022 through the Google Earth Engine (GEE) platform. To enhance data quality, cloud masking and Savitzky-Golay (SG) filtering techniques were applied to the acquired data.

Two categories of spectral data were utilized to classify soybeans and other crop types: (1) reflectance from five spectral bands and (2) the values of nine spectral indices (refer to Table  1 ). The five bands selected for classification are red edge 1 (RE1), red edge 2 (RE2), red edge 3 (RE3), shortwave infrared 1 (SWIR1) and shortwave Infrared 2 (SWIR2). The red-edge bands are important indicator bands that reflect plant pigments and health status. The short-wave infrared bands can reflect changes in moisture and other biochemical components in crop leaves 42 . Previous research has confirmed the significant role they play in distinguishing between soybeans and corn 13 , 20 , 43 . Additionally, nine commonly employed spectral indexes were computed: Enhanced Vegetation Index (EVI) 44 , Green Chlorophyll Vegetation Index (GCVI) 45 , Land Surface Water Index (LSWI) 46 , Red Edge Position Index (REPI) 47 , Red Edge Normalized Difference Vegetation Index (RENDVI) 48 , Normalized Difference Phenology Index (NDPI) 49 , and Soil-Adjusted Vegetation Index (SAVI) 50 , Optimized Soil-Adjusted Vegetation Index (OSAVI) 51 , Transformed Chlorophyll Absorption in Reflectance Index (TCARI) 52 . The use of NDVI and EVI time series is widespread for extracting temporal characteristics and phenological indicators of various crops. LSWI can effectively differentiates and classifies rice due to the heightened responsiveness of corn and soybeans to leaf and soil moisture. RENDVI and REPI, which leverage the S2 Red Edge bands, are particularly suited for estimating canopy chlorophyll II and nitrogen content. OSAVI proficiently mirrors the dynamic growth of crops while simultaneously minimize the impact of background soil 51 . There exists a high correlation between crop OSAVI and their canopy chlorophyll content, which displays significant variations throughout the growth season of the crops. In crops with high chlorophyll content, such as soybeans, corn, and rice, changes in TCARI are comparatively slow. Therefore, TCARI/OSAVI demonstrates considerable sensitivity to flux in chlorophyll content 52 . Cash crops such as peanuts, cotton, potatoes, and sunflowers, potentially outside of soybeans, are derived using TCARI/OSAVI 25 .

The availability of suitable Sentinel-2 images was limited due to frequent cloud cover and rain during the soybean growing season. This presented challenges in creating the necessary time-series spectral features for classification. To address this issue, Sentinel-1 SAR (Synthetic Aperture Radar) data was utilized to establish the required time-series spectral features for classification. Sentinel-1 is equipped with a C-band synthetic aperture radar operating at a center frequency of 5.045 GHz. It provides four imaging modes: Stripmap, Interferometric Wide swath, Extra Wide swath, and Wave modes. Sentinel-1 offers dual-polarization SAR data (HH+HV, VV+VH) and has four product specifications: RAW Level-0, SLC (Single-Look Complex), GRD (Ground Range Detected), and OCN (Ocean). Due to data storage limitations, Sentinel-1 images in GEE are accessible in the GRD format, which lacks phase information. The data in GEE undergo several pre-processing steps, which include: 1) removal of thermal noise, 2) radiometric calibration, 3) terrain correction using SRTM or ASTER DEM data, and 4) conversion of terrain-corrected backscattering coefficients to decibel values. Given the potential adverse effects of SAR active microwave imaging on image quality, this study applied Refine Lee filtering and straightforward incidence angle normalization into the processing of Sentinel-1 images.

The present study employs Sentinel-1 VV/VH dual-polarization imagery to distinguish five SAR parameters for extracting soybean features. These parameters comprise the backscattering ratio for VV and VH, denoted as \({{\rm{\sigma }}}_{{\rm{VH}}}^{{\rm{0}}}\) and \({{\rm{\sigma }}}_{{\rm{VV}}}^{{\rm{0}}}\) respectively, in addition to three combinations of polarization channels: the cross-polarization ratio ( \({{\rm{\sigma }}}_{{\rm{VH}}}^{{\rm{0}}}{{\rm{/\sigma }}}_{{\rm{VV}}}^{{\rm{0}}}\) ), the cross-polarization sum ( \({{\rm{\sigma }}}_{{\rm{VH}}}^{{\rm{0}}}{{\rm{+\sigma }}}_{{\rm{VV}}}^{{\rm{0}}}\) ), and the Radar Vegetation Index (RVI). These parameters are itemized in Table  2 .

Training and validation data

Ground survey samples, include those we collected from various provinces across different years, are denoted as yellow points in Fig.  2 and list in Table  3 . The sample locations and crop types were recorded during fieldwork using mobile Geographic Information System (GIS) devices. Post-field surveys, we conducted a visual inspection of all ground samples utilizing high-resolution images from Google Earth and Sentinel-2 RGB composite images. Any samples displaying evident errors, such as the misclassification of natural vegetation as crops, were discarded. Samples located close to roads or field boundaries were also excluded. In addition, the sample data was enhanced by using existing data products 53 .

Sample generation and migration

This paper introduces a method for generating samples that employs existing samples to facilitate the spatiotemporal migration of soybean samples, even amidst constraints in sample sizes and temporal coverage (Fig.  3 ). The strategy used in this study for generating samples involves sifting out soybean and non-soybean specimens from randomly collected cropland samples. In our study area, the primary crops grown encompass soybeans, corn, rice, wheat, and other staple crops, the planting area of which comprises up to 62% of the total cultivated land area 3 . Peanuts, rapeseed, cotton, potatoes, sunflowers, and other cash crops are also cultivated. The production of winter wheat and winter rapeseed in China accounts for more than 90% of the total wheat and rapeseed production, respectively 54 . As the growth periods of these two do not overlap with that of soybeans, they are not considered when filtering sample. To distinguish soybeans from the aforementioned non-soybean samples, our method of generating samples is categorized into three parts. Initially, the ESA WorldCover 55 is used to generate random cropland samples with unspecified crop types. Subsequently, based on the findings of Huang et al . 25 , we devised the Concave-Convexity Index (CCI), which segregates random crop samples into potential soybeans and non-soybeans based on the chlorophyll content change in the crop canopy. By charting the time series curves of band reflectivity and crop vegetation indexes, it is possible to discern accurate and reliable samples from potential ones, achieved through the analysis of the typical distribution of the area under the curve .

Generate crop samples . Different forms of land cover, including tree cover, grassland, water bodies, and buildings, are often proximate to agricultural areas. These elements may influence the integrity of the crop sample, as illustrated in Fig.  4 . In our approach, a hexagonal automated sampling technique is utilized to generate random crop points 56 , which involves scattering a substantial number of points randomly within a hexagonal grid and determine the point’s classification by appraising the proportion of land cover categories within a 50-meter buffer. If a single land cover type comprises more than 90% of the area within this buffer, it is assigned as the type for that point. To verify the accuracy of samples for uncultivated area, we performed visual analysis using high-resolution Google Earth images and Sentinel-2 data from the corresponding year.

Filter potential soybean samples . We initiated a preliminary filtering process of random crop points, aimed at identifying potential soybeans and non-soybeans. Crops’ OSAVI is highly correlated with their canopy chlorophyll content, showing significant variations as the crops grow 51 . In crops with high chlorophyll content, such as soybeans, corn, and rice, TCARI changes relatively gradually 52 . As a result, the TCARI/OSAVI trend typically displays a concave pattern in these crops 25 . Conversely, crops with low chlorophyll content, such as peanuts, cotton, potatoes and sunflowers exhibit an inverse pattern, marked by noteworthy alterations in TCARI and culminating in a convex temporal curve, as illustrated in Fig.  5 . This study leverages the TCARI/OSAVI association to evaluate the concavity or convexity of the temporal growth curves of crops by establishing the CCI for the start of growing season (SOS), peak of growing season (POS), and end of growing season (EOS) 25 . Points exhibiting a CCI ≥ 0 are marked as potential non-soybean points, while those with CCI < 0 are earmarked as possible soybean points. These points are subsequently employed in the sample selection process. The dates when the EVI temporal curve attains its peak are deemed as the POS of the crop, while the SOS and EOS are calculated using the median method 57 . The formulation for CCI calculation (Eq.  1 ) is provided below:

Confirm soybean samples . Several time series curves of band reflectivity and vegetation indexes, accurately exhibiting the growth characteristics of soybeans, as presented in Fig.  6 . The curve integration of a specific parameter represents its cumulative value throughout the entire crop growing season. Soybeans, during their peak growth phase, manifest increased dryness and greener foliage, distinguishing them from other crops 27 . Consequently, soybeans tend to exhibit elevated values for EVI and SWIR2, while their LSWI values are comparatively lower. During the peak growing season, the REPI and RENDVI for corn distinctly surpass those of soybeans 20 . Conversely, soybeans demonstrate a relatively high red edge reflectance. Accordingly, we formulated two parameter sets for soybean samples screening, which includes a high-value group (EVI, RE2, and SWIR2; see Fig.  6a–c ), and a low-value group (LSWI, RENDVI, and REPI; see in Fig.  6d,e ). As a result, soybeans can be clearly distinguished from other crops through curve integration. The integration limits align with the crop’s SOS and EOS.

Sample generation process. SOS, start of growing season; POS, peak of growing season; EOS, end of growing season; CCI, Concave-Convexity Index.

Grid-based random sample point generation with ESA WorldCover on the base map, green checkmarks indicate that a sample point of that feature type is retained, red fork markers indicate that the sample is discarded, and blue star markers indicate that cropland samples are retained for subsequent sample generation.

The band reflectivity or vegetation index of a certain crop in the peak growing season follows a one-dimensional Gaussian distribution 32 . The integral values of the aforementioned two sets of time series curves of soybeans are assumed to follow a multivariate Gaussian distribution that is connected to their dimensions. The probability density function (Eq.  2 ) is given by:

Where x is the integral vector of time series curve from ether the high-value or low-value group, μ is the mean vector of the verified soybean points, and Σ is the covariance matrix. To quantify the resemblance between random crop points and verified soybean points, we employed the Mahalanobis distance measurement. This computation is performed between the randomly selected points and a multivariate Gaussian distribution, which is composed of verified soybean points. The Mahalanobis distance gauges the deviation between data points and distributions, factoring in the correlation among different dimensions. This method mitigates the effects of varying dimensions and variances, thereby yielding a more accurate representation of the correlation between two sets of data. The Mahalanobis distance (Eq.  3 ) for a multivariate vector, with a mean of μ and a covariance matrix Σ is defined as follows:

Temporal profile of TCARI/OSAVI for ( a, b ) soybeans and ( c, d ) other crops (including peanuts, cotton, and potatoes).

Time-series vegetation index curves for soybeans and other major crops.

Figure  7 depicts the scatter distributions of soybeans, corn, and rice in three-dimensional spaces for high-value and low-value groups. Three crops clearly form distinct groups. The soybean samples are clustered in the upper-right and lower-left corners of two feature spaces. This highlights the excellent ability of the area under the curve employed in this study to distinguish soybeans from other crops. In the previous sections, we computed two sets of Mahalanobis distances between filtered points and points representing soybeans from ground survey. We hypothesize that shorter distances are indicative of soybeans, therefore emphasising the need to determine the categorization threshold. The threshold is determined by analysing the multivariate Gaussian distributions using soybean samples as the basis. Points with a high probability density are indicative of the most salient features of soybeans, and these points are concentrated at the centre of the distribution. To identify these robust soybean points, we employed the Monte Carlo method to calculate the probability density p 50 at which the cumulative probability of the multivariate Gaussian distribution (which is symmetric around the centre) hits 50%. The sought-after points are those soybean points with a probability density higher than p 50 . The formula (Eqs.  4 , 5 ) for p 50 is given as:

Spatial Distribution of Soybean Features, ( a ) High-Value Group, ( b ) Low-Value Group.

Subsequently, we computed the Mahalanobis distances from robust soybean points to the multivariate Gaussian distributions of high-value group and low-value group. The 90th percentile of these distances was selected as the threshold for confirming reliable soybean points from potential soybean, namely \({D}_{Soy-Thresh-High}\) and \({D}_{Soy-Thresh-Low}\) . For non-soybean points, we determined the 95th percentile of Mahalanobis distances from all ground survey soybean samples to the distributions, establishing \({D}_{NonSoy-Thresh-High}\) and as filtering thresholds. The filtering targets include potential soybean and non-soybean. Ultimately, the criteria (Eqs.  6 , 7 ) for selecting soybean and non-soybean points from random points are as follows:

Where D High and D Low denote the Mahalanobis distances between arbitrary points and the high-value and low-value classes of soybean ground surveys, respectively. Figure  7 depicts the ellipsoidal clustering pattern of soybean points within a three-dimensional feature space. This study shows that the robust soybean points extracted through probability density occupy the central region of the ellipsoid. The percentiles of Mahalanobis distance from these points to the distribution guarantee that the filtered soybean points are also located within the ellipsoid, thereby assuring their accuracy and reliability. It is noteworthy that the computation of Mahalanobis distance involves dimension independence and standardization, which convert the ellipsoid into a sphere in the Mahalanobis distance space, thereby expanding the precision of the point filtering process.

In terms of migration strategy, we categorized the strategy into three types based on the spatiotemporal relationship between ground survey samples and migrated ones. These categories include: (1) Migrating samples from different regions in the same year to the target area (spatial migration); (2) Migrating samples from different years in the same region to the target year (temporal migration); (3) Migrating samples from different years and regions to the target year in the intended region (spatiotemporal migration). We give priority to using ground survey samples from the same region for migration (temporal migration), because soybean in the same area have relatively little inter-annual changes in planting habits, varieties, and soil texture 58 . Therefore, we think the temporal heterogeneity of soybeans in spectral parameters is smaller than the spatial heterogeneity. If there are no ground surveys in a certain area, samples from similar climate zones will be used for migration, with priority given to those from the same years, then adjacent years. Figure  8 describes the provinces and years to which the ground survey samples used for sample migration belong.

Ground survey samples used for samples migration in each province and year.

To evaluate the effectiveness of sample generation, we employed the GLAD maize and soybean map (GLAD) 13 to assess the accuracy of the generated soybean and non-soybean samples. Since the timeline of GLAD is constrained to 2019, we performed experiments and computed accuracies only for the sample points generated in 2019, implying all target years are set as 2019.

Features selection and classification

We utilized soybean phenological characteristics and spectral indexes to differentiate between soybean and non-soybean crops, as these are efficient in capturing the seasonal fluctuations in surface spectra. To compile these indicators, we annually selected data from April 1 st to November 15 th , considering the crop calendars of various regions. Table  4 presents the candidate features we employed. Statistical features of five reflectance bands - RE1-3, SWIR1, and SWIR2 - were analysed during the growing season (DOY: 90–318). These encompass the minimum, maximum, and standard deviation, as well as the 15th, 50th, and 90th percentiles. Phenological parameters obtained from the EVI time series, harmonic fitting parameters 20 , 59 , and accumulative biomass attributes 60 , 61 are also taken into account. We utilized harmonic fitting (discrete Fourier transform, Eq.  8 ) analysis to the original effective observational data to extract time-series curves, as demonstrated in the following formula:

Where f ( t ) represents fitted vegetation index value at the time instance t . The constant term is represented as a , while b corresponds to the coefficient of the first-order term. M signifies the quantity of harmonic components, and C and D stand for the coefficients of cosine and sine functions, respectively. The variable ω is the reciprocal of the number of days in a year (1/365), t represents a specific day within a year as denoted by the DOY, and e corresponds to the residual value. For temporal feature extraction, phase and amplitude are utilized with amplitude defined as the magnitude of a two-dimensional vector [ C M , D M ], and phase as the angle of the same two-dimensional vector [ C M , D M ].

In assessing biomass using EVI, the EVI is systematically calculated for every time series data point throughout the growth season. The subsequent step involves aggregating the EVI values over unique time intervals to derive the cumulative biomass features, a crucial factor for soybean classification. To accommodate potential inconsistencies stemming from diverse pixel observation frequencies in regions with incomplete image data, a trilinear interpolation approach is incorporated in this study. This method effectively corrects missing data points, thereby ensuring a uniform computation of cumulative biomass features. The formula for trilinear interpolation (Eq.  9 ) is provided below:

Within this context, y 0 denotes the sought-after interpolation outcome, with y 2 denoting known function values. Here, x 0 designates the point slated for interpolation, whereas x 1 , x 2 , and x 3 stand as the abscissas for three established reference points. The notation f [ x 1 , x 2 , x 3 ] represents the third-order divided difference calculated at positions x 1 , x 2 , and x 3 .

In addressing the five SAR feature parameters (Table  5 ), we harnessed SAR image time series to extract pivotal phenological characteristics of crops. This process comprised both statistical features and principal component features. The statistical features mirrored the approach adopt for optical data, incorporating the maximum, minimum, and variance, along with the 15th, 50th, and 90th percentiles of the five SAR parameters. These statistical attributes are instrumental in conveying the average levels and temporal fluctuations within time series curves for diverse crops. Furthermore, we carried out a Principal Component Analysis (PCA) on the Sentinel-1 image time series in the temporal domain, selecting the initial three principal components as the principal component features of SAR data 62 .

Considering the ‘Hughes’ Phenomenon, the current number of features is copious. Consequently, in each classification process, we incorporate feature selection, choosing to maintain the top 50% of features. This decision is influenced by the ranking provided by the random forest model for the final classification.

We employed local random forest classifiers for soybean planting areas identification in each province. This non-parametric machine learning classifier exhibits a higher error tolerance compared to certain parametric classifiers and has been extensively utilized in classification and recognition research. In terms of dividing training set and testing set, half of the samples within each province were randomly selected for training the classifier and mapping soybean cropland, while the remainder were utilized for validation. In this study, we implement the random forest classification model within the GEE. On the GEE platform, we vary the number of decision trees from 50 to 500 at 50-unit intervals. The chosen number of decision trees as the parameter for ensuring classification is the one that surpasses 100 and achieves the initial local maximum in classification accuracy. To counteract minor result variations in each experimental repetition due to the inherent randomness in random forest sampling, we set a random seed of 999. All additional parameters are left at their default values.

To assess the precision of soybean distribution mapping, we take two approaches: (1) on-site validation through the collection of ground truth samples, which involves conducting ground surveys and generating samples, and (2) comparing the results with agricultural statistical data obtained from administrative units. Confusion matrices were generated using both soybean samples and non-soybean samples for each provincial soybean map. These matrices were employed to calculate the producer’s accuracy (PA), user’s accuracy (UA) and F1-score (F1) for soybean samples (Eqs.  10 – 12 ), assessing the precision of the approaches. The overall success of this strategy was assessed by calculating the overall accuracy (OA). The Kappa coefficient was used to assess the level of agreement between the classification results and sample labels. In addition, we assessed the soybean planting area identified in this study by comparing it to agricultural statistical data at the provincial and prefectural levels. This comparison was done using the coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE).

Post-processing

For large-scale and high-resolution crop mapping, the speckle noise is inevitable, and the same goes for soybean mapping 63 . Errors may arise during sensor imaging, soybean sample generation, image preprocessing, and feature classification, etc., resulting in soybean patches composed of just one or two pixels in the mapping results. In most cases, they are considered speckle noise and should be eliminated. We performed post-processing on the results using eight-neighborhood majority filtering. This processing can filter independent, unconnected soybean pixels, and non-soybean pixels in soybean plots will also be filled, making the mapping results more accurate and reasonable.

Data Records

Between 2019 and 2022, we generated four soybean cropland maps encompassing China’s key soybean-producing regions, all at a 10-meter spatial resolution (ChinaSoybean10). The datasets, formatted to Geotiff, are available for access at the Zenodo repository ( https://doi.org/10.5281/zenodo.10068402 ) 64 . Structured under the ESPG: 4326 (WGS_1984) spatial reference system, the maps incorporate only one values: 1 to denote soybean planting areas, and null value to indicate non-soybean planting areas (inclusive of other landcover). These maps can be scrutinized and visualized using software such as ArcGIS, QGIS, or their alternatives.

Technical Validation

Precision assessment of sample spatiotemporal migration.

Employing GLAD maize and soybean map 13 , we evaluated the sample generation accuracy for both soybeans and non-soybeans. GLAD maize and soybean map ( https://glad.earthengine.app/view/china-crop-map ) is a 2019 national maize and soybean map produced using field survey samples and binary random forest, in which the R 2 between soybean mapping area and statistical yearbook area can reach 0.93. It is considered to be a reliable reference for accuracy validation. Using the actual samples from 2019 to 2021, we generated soybean and non-soybean samples for different regions in 2019 via three different methods. Subsequently, we calculated the sample generation accuracy for each region, as delineated in Fig.  9 . Broadly speaking, with the exception of Liaoning and Jiangsu, the generation accuracy exceeds 80% for soybean samples and 95% for non-soybean samples, indicating the efficacy of the method used. Among the three-generation methods applied for soybean samples–temporal migration, spatial migration, and spatiotemporal migration–the average accuracies were 87.32%, 86.49%, and 83.44%, respectively. Temporal migration within the same region proved to be superior, followed by spatial migration. The least accuracy occurs in spatiotemporal migration. We postulated that minor annual variations in climatic factors, such as temperature and precipitation, contribute less to negative effects on sample migration compared to spatial heterogeneity due to regional differences. Among all provinces, Heilongjiang demonstrated the best results with an average accuracy of 92.72%. Inner Mongolia, Anhui, and Henan, three major soybean-producing provinces, also reached approximately 90%. In these provinces, soybeans are extensively cultivated, leading to relatively continuous and dense area, resulting in high accuracy. In contrast, Liaoning and Jiangsu, characterized by complex planting structures and fragmented soybean planting areas, had an accuracy below 80%. This lower accuracy can be attributed to these agricultural complexities and the adverse impact of rainfall on image quality. Conversely, the generation of non-soybean samples illustrated commendable accuracy and robustness across various regions. In summary, our sample generation method demonstrated exceptional proficiency across diverse regions, delivering commendable outcomes in both temporal and spatial sample generation.

2019 generated sample precision evaluation, ( a ) Soybeans, ( b ) Non-Soybeans, I-III represents different methods for sample migration: I-temporal migration, II-Spatial migration, III-Spatiotemporal migration.

Soybean map and accuracy assessment

By harnessing Sentinel-2 remote sensing imagery, selected Sentinel-1 SAR image data, ground surveys and generated samples, we mapped soybean planting areas for ten provinces nationwide (Fig.  10 ), denoted as ChinaSoybean10. We conducted accuracy assessments of the mapping results using both ground surveys and generated samples. The result indicates that in the northeastern region, the average overall accuracy for soybean planting areas mapping was 93.70%, with a prevailing Kappa coefficient of 0.8624. In crucial soybean cultivation areas of the Huang-Huai-Hai Plain, the middle-lower reaches of Yangtze River Plain, and Sichuan, the average overall mapping accuracy was 93.16%, accompanied by a Kappa coefficient of 0.7980 (Table  6 ). Moreover, we calculated both the producer’s accuracy and the user’s accuracy for each province. In the prominent soybean planting areas of the Northeast, the average producer’s accuracy, user’s accuracy, and F1-score were 92.23%, 88.70%, and 90.06%, respectively. For the Huang-Huai-Hai Plain, the Middle-Lower Yangtze Plain, and Sichuan, the average of these indicators were 80.15%, 89.59%, and 0.8434, respectively.

The crop maps in the main soybean producing area of China in ( a ) 2019, ( b ) 2020, ( c ) 2021, and ( d ) 2022.

We compared the mapped soybean cropland areas of various prefecture-level cities with the officially reported planting areas, and quantitatively analyzed the accuracy of our soybean map by calculating R-squared ( R 2 ), root-mean-square-error (RMSE), and mean-absolute-error (MAE). The results demonstrate a high level of consistency between our annual soybean maps and official statistical data ( R 2  > 0.85), with values of 0.91, 0.92, 0.87, and 0.88 for the years 2019 to 2022, respectively (Fig.  11 ). R 2 for 2019 and 2020 were both above 0.9, while there was a slight decrease in 2021 and 2022, likely due to the increased use of generated soybean samples in those years. In Fig.  11 , we plotted the 1:1 line, and some prefecture-level cities with higher soybean production, such as Heihe, Qiqihar, and Hulunbuir, tended to cluster around this line. However, certain cities in the Huang-Huai-Hai and Yangtze River regions exhibited slight discrepancies compared to the statistical yearbook, possibly due to the more complex planting structures and the presence of numerous smallholders 65 , 66 , resulting in significant field-level heterogeneity. Nevertheless, our method consistently produces highly reliable estimates of planting area.

Comparison of mapped soybean area and planted soybean area reported by statistics at prefectural levels in ( a ) 2019, ( b ) 2020, ( c ) 2021, ( d ) 2022.

Visual comparison with other products and methods

Compared with some existing soybean distribution products, ChinaSoybean10 has wider spatial and temporal coverage. It also merges multi-source remote sensing data to achieve superior classification accuracy. We rank ChinaSoybean10 alongside GLAD maize-soybean map 13 , CDL (Crop Data Layer) 20 and soybean map produced by GWCCI 27 . Among them, GLAD covers the soybean planting areas across the country in 2019, CDL covers the Northeast region, and GWCCI uses a common threshold of 0.17 for soybean mapping in all regions across the country. As a demonstration reference in 2019, we selected examples from Heilongjiang (Fig.  12a,b ), and Anhui (Fig.  12c ) to illustrate the comparison between our soybean mapping results and those of state-of-the-art methods. For the first example (Fig.  12a ), our soybean mapping results are in good agreement with CDL and GLAD, reflecting the accuracy of our results. In the second example (Fig.  12b ), our results are very similar to GLAD, while CDL has redundant soybean recognition results in the red-boxed area. The third example is in Anhui Province (Fig.  12c ). CDL does not cover this area, so the GWCCI generated results are compared with our results. The soybean mapping effectiveness of GWCCI relies heavily on image quality and threshold selection. It can be found that the overall result contains more salt and pepper noise, as well as misclassified pixels in the red-boxed area (Fig.  12c3 ), which do not exist in our results and GLAD. The above comparison verifies the accuracy of our results.

Comparison with Existing Methods and Results, the first column is a median-synthesized RGB image from Sentinel-2 after cloud removal (from DOY 200 to DOY 240); the second column represents ChinaSoybean10; The second column shows the results of CDL and the results extracted by GWCCI, where (c1-c2) are from CDL and (c3) is the extraction result of GWCCI.; the fourth column is the soybean map of GLAD, with ( a – c ) indicating different regions: ( a, b ) 2019 soybean map in Heilongjiang province; ( c ) 2019 soybean map in Anhui province.

Advantages of the sample migration method

The paper presents a soybean sample migration method, optimizing the acquisition of crop samples for soybean area mapping. The method is less financially demanding, automated, and provides accurate soybean and non-soybean samples using minimal pre-existing ones. Figure  13 displays samples taken in 10 provinces during 2019. The spatiotemporal migration is based on the disparity in crop’s band reflectivity and vegetation indexes. Using the integration of this curve during the growing season improves migration accuracy. The method’s potential lies in reducing the spatiotemporal heterogeneity of soybean phenology by using automatically determined growth season intervals.

2019 generated soybean sample and ground surveys in 10 provinces. ( a ) Heilongjiang, ( b ) Inner Mongolia, ( c ) Jilin, ( d ) Liaoning, ( e ) Anhui, ( f ) Henan, ( g ) Shandong, ( h ) Hubei, ( i ) Jiangsu, ( j ) Sichuan.

Currently, there has been considerable research on no sample crop mapping based on crop knowledge and rule thresholds 25 , 27 , 32 . However, the effectiveness of these methods is severely hampered by threshold selection and are susceptible to clouds and fog, thus posing difficulties for large-scale crop mapping. The method presented in this paper selects random crop points based on the feature distribution of soybean samples, automatically determining filtering thresholds through the distribution characteristics of ground survey samples, thus enhancing its generality. Additionally, the mapping strategy in this paper combines generated samples with supervised classification. The thresholds mainly restrict the quality and quantity of samples, rather than directly affecting the mapping results. Therefore, threshold calculation parameters can be more “extreme” to obtain a purer set of soybean samples. Compared to the sample migration method based on DTW distance developed by Zhang et al . 33 , our method eliminates the need to obtain samples of other major crops in the target area, hence offering a more versatile and convenient solution. Some researchers use crop samples generated by existing products for mapping 13 , 56 . Although this method is convenient and efficient, it is subject to considerable limitations in terms of both time and space.

In conclusion, the soybean sample migration methodology elucidated in this paper adeptly and efficiently procures both soybean and non-soybean samples for regions devoid of such samples. This significantly aids in the creation of comprehensive crop mapping products and offers myriad possibilities for crop mapping.

Usage Notes

China is the fourth largest soybean producer and the largest soybean importer in the world, and its soybean consumption relies heavily on imports. Mapping the distribution of soybean growing areas at the national scale is critical for food and energy security in the context of growing population and consumption. In this paper, we collected soybean field survey samples for many years and proposed a sample spatiotemporal migration method based on the temporal characteristics of vegetation index. Using field survey and generated samples, we create national 10 m soybean maps in China from 2019 to 2022. Through experiments, we found that the areas calculated by our soybean maps area consistent highly with the official statistical area at the prefecture-level. Therefore, our soybean mapping results can be used to support large-scale soybean yield estimates and quantitative analyzes of multi-year soybean-cultivated area changes. Furthermore, our datasets can serve as a reference and support uncertainty analysis for comparable products.

Uncertainty

Despite our stringent data processing measures, certain sources of uncertainty remain inherent. Though we employed time-series Sentinel-2 data for soybean planting area mapping, the length of the soybean growing season is geographically variable. The 5-day revisit cycle might not consistently yield complete time-series spectral curves due to obstacles such as cloudy conditions and rainfall. While we successfully integrated Sentinel-1 data in certain regions and years, completely eradicating the speckle noise remains a complex task. Further, the widespread practice of soybean intercropping with crops such as corn and sorghum presents a substantial challenge in accurately mapping soybean’s spatial distribution within the Huang-Huai-Hai and Yangtze River regions. The potential mixed pixel effect arising from the 10-meter spatial resolution of Sentinel-2 data inevitably weakens the identification signal of specific crops, introducing uncertainty. Lastly, despite the validation of our sample generation method using ground-truth samples and existing products by achieving approximately 90% sample accuracy, potential deviations may still affect our results. Additionally, our methodology still relies upon terrestrial survey soybean samples. Ensuring minor phenological differences between these field survey samples, and generated samples is critical to the accuracy of sample generation. Looking forward, leveraging crop-specific maps, and highly resolved remote sensing products could offer solutions for the mixed pixel issue and enhance sample generation methods for optimal differentiation between soybean and non-soybean areas. Consequently, this could simplify the soybean extraction process.

Code availability

The programs used to generate the datasets and all the results were ESRI ArcGIS (10.6), Python (3.7 or 3.8) and Google Earth Engine (GEE). The scripts utilized for ChinaSoybean10 described in this paper can be accessed at https://github.com/ZihangLou/ChinaSoybean10 .

He, L., Jin, N. & Yu, Q. Impacts of climate change and crop management practices on soybean phenology changes in China. Sci. Total Environ. 707 , 135638 (2020).

Article   ADS   CAS   PubMed   Google Scholar  

Yao, G., Hertel, T. W. & Taheripour, F. Understanding China’s Soybean Boom from Historical Validation. (2017).

National Bureau of Statistics of China. China Statistical Yearbook. (2022).

Liu, Z. et al . Optimization of China’s maize and soy production can ensure feed sufficiency at lower nitrogen and carbon footprints. Nat. Food 2 , 426–433 (2021).

Article   PubMed   Google Scholar  

Najafi, E., Devineni, N., Khanbilvardi, R. M. & Kogan, F. Understanding the Changes in Global Crop Yields Through Changes in Climate and Technology. Earths Future 6 , 410–427 (2018).

Article   ADS   CAS   Google Scholar  

Bégué, A. et al . Remote Sensing and Cropping Practices: A Review. Remote Sens. 10 , 99 (2018).

Article   ADS   Google Scholar  

Weiss, M., Jacob, F. & Duveiller, G. Remote sensing for agricultural applications: A meta-review. Remote Sens. Environ. 236 , 111402 (2020).

Article   Google Scholar  

Zhang, H., Kang, J., Xu, X. & Zhang, L. Accessing the temporal and spectral features in crop type mapping using multi-temporal Sentinel-2 imagery: A case study of Yi’an County, Heilongjiang province, China. Comput. Electron. Agric. 176 , 105618 (2020).

Liu, X. et al . Geographic information science in the era of geospatial big data: A cyberspace perspective. The Innovation 3 , 100279 (2022).

Article   PubMed   PubMed Central   Google Scholar  

Boryan, C., Yang, Z., Mueller, R. & Craig, M. Monitoring US agriculture: the US Department of Agriculture, National Agricultural Statistics Service, Cropland Data Layer Program. Geocarto Int. 26 , 341–358 (2011).

Wang, S., Tommaso, D., Deines, S., Lobell, J. M. & Mapping, D. B. twenty years of corn and soybean across the US Midwest using the Landsat archive. Sci. Data 7 , 307 (2020).

Peng, Q. et al . A twenty-year dataset of high-resolution maize distribution in China. Sci. Data 10 , 658 (2023).

Li, H. et al . Development of a 10-m resolution maize and soybean map over China: Matching satellite-based crop classification with sample-based area estimation. Remote Sens. Environ. 294 , 113623 (2023).

Turker, M. & Arikan, M. Sequential masking classification of multi‐temporal Landsat7 ETM+ images for field‐based crop mapping in Karacabey, Turkey. Int. J. Remote Sens. 26 , 3813–3830 (2005).

Vescovi, F. D. & Gomarasca, M. A. Integration of Optical and Microwave Remote Sensing Data for Agricultural Land Use Classification. Environ. Monit. Assess. 58 , 133–149 (1999).

Jia, K., Wu, B. & Li, Q. Crop classification using HJ satellite multispectral data in the North China Plain. J. Appl. Remote Sens. 7 , 073576 (2013).

Kpienbaareh, D. et al . Crop Type and Land Cover Mapping in Northern Malawi Using the Integration of Sentinel-1, Sentinel-2, and PlanetScope Satellite Data. Remote Sens. 13 , 700 (2021).

Kumari, M., Pandey, V., Choudhary, K. K. & Murthy, C. S. Object-based machine learning approach for soybean mapping using temporal sentinel-1/sentinel-2 data. Geocarto Int. 37 , 6848–6866 (2022).

Tran, K. H., Zhang, H. K., McMaine, J. T., Zhang, X. & Luo, D. 10 m crop type mapping using Sentinel-2 reflectance and 30 m cropland data layer product. Int. J. Appl. Earth Obs. Geoinformation 107 , 102692 (2022).

You, N. et al . The 10-m crop type maps in Northeast China during 2017–2019. Sci. Data 8 , 41 (2021).

Han, J. et al . The RapeseedMap10 database: annual maps of rapeseed at a spatial resolution of 10&thinsp;m based on multi-source data. Earth Syst. Sci. Data 13 , 2857–2874 (2021).

You, N. & Dong, J. Examining earliest identifiable timing of crops using all available Sentinel 1/2 imagery and Google Earth Engine. ISPRS J. Photogramm. Remote Sens. 161 , 109–123 (2020).

Zhu, M. et al . Identification of soybean based on Sentinel-1/2 SAR and MSI imagery under a complex planting structure. Ecol. Inform. 72 , 101825 (2022).

Ren, T., Xu, H., Cai, X., Yu, S. & Qi, J. Smallholder Crop Type Mapping and Rotation Monitoring in Mountainous Areas with Sentinel-1/2 Imagery. Remote Sens. 14 , 566 (2022).

Huang, Y. et al . Automated soybean mapping based on canopy water content and chlorophyll content using Sentinel-2 images. Int. J. Appl. Earth Obs. Geoinformation 109 , 102801 (2022).

Xu, J. et al . Towards interpreting multi-temporal deep learning models in crop mapping. Remote Sens. Environ. 264 , 112599 (2021).

Chen, H. et al . A novel Greenness and Water Content Composite Index (GWCCI) for soybean mapping from single remotely sensed multispectral images. Remote Sens. Environ. 295 , 113679 (2023).

Zhong, L., Hu, L., Yu, L., Gong, P. & Biging, G. S. Automated mapping of soybean and corn using phenology. ISPRS J. Photogramm. Remote Sens. 119 , 151–164 (2016).

Luo, Y., Zhang, Z., Zhang, L. & Cao, J. Spatiotemporal patterns of winter wheat phenology and its climatic drivers based on an improved pDSSAT model. Sci. China Earth Sci. 64 , 2144–2160 (2021).

Maxwell, A. E., Warner, T. A. & Fang, F. Implementation of machine-learning classification in remote sensing: an applied review. Int. J. Remote Sens. 39 , 2784–2817 (2018).

Lin, C. et al . Early- and in-season crop type mapping without current-year ground truth: Generating labels from historical information via a topology-based approach. Remote Sens. Environ. 274 , 112994 (2022).

You, N., Dong, J., Li, J., Huang, J. & Jin, Z. Rapid early-season maize mapping without crop labels. Remote Sens. Environ. 290 , 113496 (2023).

Zhang, S. et al . Crop type mapping with temporal sample migration. Int. J. Remote Sens . 1–19 https://doi.org/10.1080/01431161.2023.2192881 (2023).

Rusňák, T. et al . Crop Mapping without Labels: Investigating Temporal and Spatial Transferability of Crop Classification Models Using a 5-Year Sentinel-2 Series and Machine Learning. Remote Sens. 15 , 3414 (2023).

Zhong, L., Gong, P. & Biging, G. S. Efficient corn and soybean mapping with temporal extendability: A multi-year experiment using Landsat imagery. Remote Sens. Environ. 140 , 1–13 (2014).

Belgiu, M., Bijker, W., Csillik, O. & Stein, A. Phenology-based sample generation for supervised crop type classification. Int. J. Appl. Earth Obs. Geoinformation 95 , 102264 (2021).

Song, X.-P. et al . Massive soybean expansion in South America since 2000 and implications for conservation. Nat. Sustain. 4 , 784–792 (2021).

Di, Y. et al . Recent soybean subsidy policy did not revitalize but stabilize the soybean planting areas in Northeast China. Eur. J. Agron. 147 , 126841 (2023).

Zhao, L. et al . In-season crop type identification using optimal feature knowledge graph. ISPRS J. Photogramm. REMOTE Sens. 194 , 250–266 (2022).

Liu, W. et al . Farmland Parcel Mapping in Mountain Areas Using Time-Series SAR Data and VHR Optical Images. Remote Sens. 12 , 3733 (2020).

Burke, M. & Lobell, D. B. Satellite-based assessment of yield variation and its determinants in smallholder African systems. Proc. Natl. Acad. Sci. 114 , 2189–2194 (2017).

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Luo, C. et al . Monthly composites from Sentinel-1 and Sentinel-2 images for regional major crop mapping with Google. Earth Engine. J. Integr. Agric. 20 , 1944–1957 (2021).

Segarra, J., Buchaillot, M. L., Araus, J. L. & Kefauver, S. C. Remote Sensing for Precision Agriculture: Sentinel-2 Improved Features and Applications. Agronomy 10 , 641 (2020).

Article   CAS   Google Scholar  

Huete, A. et al . Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sens. Environ. 83 , 195–213 (2002).

Gitelson, A. A. et al . Remote estimation of leaf area index and green leaf biomass in maize canopies. Geophys. Res. Lett . 30 , (2003).

Xiao, X., Boles, S., Liu, J., Zhuang, D. & Liu, M. Characterization of forest types in Northeastern China, using multi-temporal SPOT-4 VEGETATION sensor data. Remote Sens. Environ. 82 , 335–348 (2002).

Frampton, W. J., Dash, J., Watmough, G. & Milton, E. J. Evaluating the capabilities of Sentinel-2 for quantitative estimation of biophysical variables in vegetation. ISPRS J. Photogramm. Remote Sens. 82 , 83–92 (2013).

Defourny, P. et al . Near real-time agriculture monitoring at national scale at parcel resolution: Performance assessment of the Sen2-Agri automated system in various cropping systems around the world. Remote Sens. Environ. 221 , 551–568 (2019).

Zhang, X. et al . Monitoring vegetation phenology using MODIS. Remote Sens. Environ. 84 , 471–475 (2003).

Huete, A. R. A soil-adjusted vegetation index (SAVI). Remote Sens. Environ. 25 , 295–309 (1988).

Rondeaux, G., Steven, M. & Baret, F. Optimization of soil-adjusted vegetation indices. Remote Sens. Environ. 55 , 95–107 (1996).

Haboudane, D., Miller, J. R., Tremblay, N., Zarco-Tejada, P. J. & Dextraze, L. Integrated narrow-band vegetation indices for prediction of crop chlorophyll content for application to precision agriculture. Remote Sens. Environ. 81 , 416–426 (2002).

Zhi, F. et al . Rapid and Automated Mapping of Crop Type in Jilin Province Using Historical Crop Labels and the Google Earth Engine. Remote Sens. 14 , 4028 (2022).

Tao, J., Zhang, X., Wu, Q. & Wang, Y. Mapping winter rapeseed in South China using Sentinel-2 data based on a novel separability index.  J. Integr. Agric. 22 , 1645–1657 (2023).

Zanaga, D. et al . ESA WorldCover 10 m 2020 V100. 2021 . (2021).

Xuan, F. et al . Mapping crop type in Northeast China during 2013–2021 using automatic sampling and tile-based image classification. Int. J. Appl. Earth Obs. Geoinformation 117 , 103178 (2023).

Diao, C. Remote sensing phenological monitoring framework to characterize corn and soybean physiological growing stages. Remote Sens. Environ. 248 , 111960 (2020).

Falco, N. et al . Influence of soil heterogeneity on soybean plant development and crop yield evaluated using time-series of UAV and ground-based geophysical imagery. Sci. Rep. 11 , 7046 (2021).

Zhang, J., Feng, L. & Yao, F. Improved maize cultivated area estimation over a large scale combining MODIS–EVI time series data and crop phenological information. ISPRS J. Photogramm. Remote Sens. 94 , 102–113 (2014).

Shammi, S. A. & Meng, Q. Use time series NDVI and EVI to develop dynamic crop growth metrics for yield modeling. Ecol. Indic. 121 , 107124 (2021).

Peroni Venancio, L. et al . Potential of using spectral vegetation indices for corn green biomass estimation based on their relationship with the photosynthetic vegetation sub-pixel fraction. Agric. Water Manag. 236 , 106155 (2020).

Mashaba-Munghemezulu, Z., Chirima, G. J. & Munghemezulu, C. Mapping Smallholder Maize Farms Using Multi-Temporal Sentinel-1 Data in Support of the Sustainable Development Goals. Remote Sens. 13 , 1666 (2021).

Li, X. et al . Mapping annual 10-m maize cropland changes in China during 2017–2021. Sci. Data 10 , 765 (2023).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Zhang, H. et al . ChinaSoybean10:An Annual 10-m Soybean cropland Mapping Dataset in China from 2019 to 2022. Zenodo https://doi.org/10.5281/zenodo.10068402 (2024).

Zhao, Y., Chen, X. & Lobell, D. B. An approach to understanding persistent yield variation—A case study in North China Plain. Eur. J. Agron. 77 , 10–19 (2016).

Zhang, W. et al . Closing yield gaps in China by empowering smallholder farmers. Nature 537 , 671–674 (2016).

Pageot, Y., Baup, F., Inglada, J., Baghdadi, N. & Demarez, V. Detection of Irrigated and Rainfed Crops in Temperate Areas Using Sentinel-1 and Sentinel-2 Time Series. Remote Sens. 12 , 3044 (2020).

Khabbazan, S. et al . Crop Monitoring Using Sentinel-1 Data: A Case Study from The Netherlands. Remote Sens. 11 , 1887 (2019).

Kaplan, G. et al . Normalizing the Local Incidence Angle in Sentinel-1 Imagery to Improve Leaf Area Index, Vegetation Height, and Crop Coefficient Estimations. Land 10 , 680 (2021).

Chang, J. G., Shoshany, M. & Oh, Y. Polarimetric Radar Vegetation Index for Biomass Estimation in Desert Fringe Ecosystems. IEEE Trans. Geosci. Remote Sens. 56 , 7102–7108 (2018).

Download references

Acknowledgements

The Strategic Priority Research Program of the Chinese Academy of Sciences (XDA28050100), this work was partially supported by the National Key Research and Development Program of China (2019YFE0115200), and the National Natural Science Foundation of China (42071329). The authors also would like to thank the anonymous reviewers for their thoughtful comments and efforts towards improving our manuscript.

Author information

These authors contributed equally: Hongchi Zhang, Zihang Lou.

Authors and Affiliations

Key Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, 100094, China

Hongchi Zhang, Zihang Lou, Dailiang Peng, Bing Zhang, Jinkang Hu, Songlin Yang & Enhui Cheng

International Research Center of Big Data for Sustainable Development Goals, Beijing, 100094, China

Hongchi Zhang, Zihang Lou, Dailiang Peng, Jinkang Hu, Songlin Yang & Enhui Cheng

University of Chinese Academy of Sciences, Beijing, 100094, China

Hongchi Zhang, Zihang Lou, Bing Zhang, Jinkang Hu, Songlin Yang & Enhui Cheng

Jiangxi Nuclearindustry Surveying and Mapping Institute Group Co., Ltd, Nanchang, 330038, China

College of Land Science and Technology, China Agricultural University, Beijing, 100083, China

Jianxi Huang

Geospatial Sciences Center of Excellence, Department of Geography Geospatial Sciences, South Dakota State University, Brookings, SD, 57007, USA

Xiaoyang Zhang

Department of Earth System Science, Tsinghua University, Beijing, 100084, China

Institute of Applied Remote Sensing & Information Technology, Zhejiang University, Hangzhou, 310058, China

National Engineering Research Center for Agro-Ecological Big Data Analysis & Application, Anhui University, Hefei, 230601, China

Linsheng Huang

Innovation Academy for Microsatellites, Chinese Academy of Sciences, Shanghai, 200120, China

Guohua Liu & Shuang Gao

You can also search for this author in PubMed   Google Scholar

Contributions

D.P. and B.Z. designed the research and developed the methodology; H.Z., Z.L. and W.L. developed the workflow and conducted the construction of the dataset; D.P., B.Z., J.X.H., J.K.H., L.H. and E.C. collected datasets; H.Z. and Z.L. drafted the manuscript; D.P. and B.Z. edited and revised the paper. All authors contributed to the interpretation of the results, provided in-depth advice, and commented/edited the manuscript. H.Z. and Z.L. contributed equally to this work and should be considered co-first author.

Corresponding authors

Correspondence to Dailiang Peng or Bing Zhang .

Ethics declarations

Competing interests.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Zhang, H., Lou, Z., Peng, D. et al. Mapping annual 10-m soybean cropland with spatiotemporal sample migration. Sci Data 11 , 439 (2024). https://doi.org/10.1038/s41597-024-03273-5

Download citation

Received : 22 December 2023

Accepted : 16 April 2024

Published : 02 May 2024

DOI : https://doi.org/10.1038/s41597-024-03273-5

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Anthropocene newsletter — what matters in anthropocene research, free to your inbox weekly.