Application of remote sensing technology in crop acreage and yield statistical survey in China

Zhang Fumina, Zhu Zaichunb, Pan Yaozhongb, Hu Tangaob, Zhang Jinshuib

a Computer Center of National Bureau of Statistics of China, Beijing 100826, China;

b State Key Laboratory of Earth Surface Processes and Resource Ecology, College of Resources

Science & Technology, Beijing Normal University, Beijing 100875, China

Contents

1. Development process of crop area and yield estimation survey system in China

2. Development of crop area and yield estimation by remote sensing measurements

3. Application status of agriculture survey by remote sensing in China

4. Summary and outlook

1. Development process of crop area and yield estimation survey system in China

1.1 Historical origin

Development process of the existing survey system in China can

be divided into five phases:

(1) The initial stage in 50s, 20th century;

(2) The stage of development in 60s, 20th century;

(3) The early recovery period in 80s, 20th century;

(4) The full implementation phase after 1984;

(5) Sample rotation and multi-subject survey after 2000.

1.2 Development needs

The current survey system has following shortages:

(1)foundation of sample survey is not solid enough;

(2)sample survey objects are unstable;

(3)sampling methods are backward;

(4)the quality of investigation is lack of control.

1.2 Development needs

we need to further improve the crop sampling system from the

following aspects:

(1) Improve the survey system: from the traditional catalog sample to space sample,

and expand the connotation of the investigation targets, provide not only crop area and

yield in digital form, but also with crop spatial distribution information and survey result

that can be used for provincial, county and village

(2) Improve the survey results: reducing the impact of subjective factors, improve the

objectivity of the results, enhance the timeliness of reports, and dynamically monitor the

crops in the critical period of growing season;

(3) Improve means of investigation: use the "3S" techniques to replace the existing

tools such as ground measuring using the tape.

2. Development of crop area and yield estimation by remote sensing measurements

2.1 foreign major research programs and technical systems

Date Project Country

1974-1978 LACIE The United States

1980-1986 AGRISTARS The United States

1989-1993 MARS The European Union

2001-? LUCAS The European Union

1980s-Now et al. et al.

2.2 Development progress of major research projects and technical system in China

Date Project

“seventh five year-plan”(1986-1990)

The Chinese Meteorological Administration : Crop monitoring by remote sensing

“eighth five year-plan”(1991-1995)

The Chinese Academy of Sciences: Crop monitoring by remote sensing

“ninth five year-plan”(1996-2000)

The Ministry of Agriculture : Crop monitoring by remote sensing

“tenth five year-plan”(2001-2005)

1 Agricultural condition monitoring system of The Ministry of Agriculture2 GVG system3 Multi-user remote-sensing business operations sampling techniques and systems to achieve4 Application and technology standards of crop area measurement by remote sensing

“eleventh five year-plan”(2006-2010)

NSRCP

3. National Statistics and Remote Sensing System of Crop Production(NSRCP)

3.1 Framework of NSRCP

Data

Flow

Data

Management

Data

Transmission

Results

Data Analysis and

Processing

Data Management and Sharing Platform（Including PDA Survey System）Data Ordering System Standardization System Standardization

Store

Obtain

Basic Data Ordering

Basic Subject Data Ordering

Phenological Data Ordering

Meteorological Station Data Ordering

RS Image Ordering

Plot Survey Data Ordering

Basic Data Standardization

Basic Subject Data Standardization

Phenological Data Standardization

Meteorological Data Standardization

RS Image Standardization

Sample Standardization

Crop Acreage Measurement System

Growth and Yield Measurement System

Hardware Network Platform

Database Platform

Basic Software Platform

Server Storage Array Network SystemGround Investigation

Network…

RS Software GIS Software Database Software Statistical Software …

Basic Database Subject Database Present RS Image DatabasePresent Statistical Survey

DatabaseStatistics and Remote Sensing

product Database

Sampling Survey System

Ground Sampling System

Spatial Sampling System

Product Publish System

Producing

Production Analysis

Production Publish

RS Image Measurement

Yield Parameters Inversion

Growth Monitor Model

Yield Estimation Model

Production Testing

Ground Sampling Production Testing

Spatial Sampling Production Testing

Crop Acreage Measurement ResultSampling Result

Yield E

stimation R

esultNew Sensors Acreage

Measurement

Acreage Preliminary Measurement

Spatial Optimization

Acreage Production Testing

Statistical design(Data Ordering and Standardization)

Business Flow

Statistical Survey(Data Analysis and Processing)

Statistical Collation and Report

Statistical product publish

Planting Structure Consult System

Regular Field Sam

plesField Sam

ples

Real Y

ield Survey

Framework of NSRCP

3.2 Survey System of NSRCP

Survey crops : wheat, corn and rice.

Survey contents: Acreage and Yield.

Accuracy : better than 95% under the confidence level of 95% at

the county-level.

Report frequency : Acreage should be annually report, and the

yield should be report three times: preliminary report, pre-

production and actual production.

3.3 Key technology system of NSRCP

There are four key technology system of NSRCP:

(1)Geo-spatial framework

(2)Ground-support network

(3)Planting area measurement

(4)Growth monitoring and yield estimation

3.3.1 Geo-spatial framework

Geo-spatial framework is the basis of NSRCP.

Geo-spatial framework mainly includes:

1 Theoretical framework construction

2 Database system design

3 Application demonstration of space-based statistics and

remote sensing and so on.

3.3.1 Geo-spatial framework

Software Hardware Network

Raster Map Vectorization

Table Spatialization RS Image Serialization

Data Cleansing,Intergration and Database Construction

Multi Source RS Image Base Map Framework

Agricultural Statistical Data Framework

GIS Data Framework

Statistics and Remote Sensing Framework Database System

Geo-spatial Framework Management Platform

Geo-spatial Framework Publish Platform

Geo-spatial Framework Database

System Management Platform

System Publish Platform

Demonstration Systems in NSRCP

Supportive Deposition

Data Gathering and Updating

Framework Data Construction

Software Development

Demonstration Application System

Policies And Standards of Geo-spatial framework

Data Preprocessing

Geo-spatial framework of NSRCP

3.3.2 Ground-support network

Ground-support network achieve associated ground survey information of planting area and production of main crops timely and accurately.

The main survey content includes: sampling plot survey, high resolution area survey, typical area survey, county-level bio-meteorology survey and cropping structure of sampling village survey.

3.3.2 Ground-support network

Sampling Block

Planting Structure

Plots with High Resolution Image

Typical Area

Phenology at County Level

Ground survey organization chart

3.3.3 Planting area measurement

In support of mass to-ground sampling data by National Bureau of Statistics, crop acreage measurement methodology mainly includes:

1 Regionalization of crop acreage measurement

2 Measurement techniques at regional/county level (remote sensing)

3 Measurement techniques and methods at the county level (remote sensing and sampling)

4 Measurement techniques and methods at the provincial level (remote sensing and sampling)

5 Analysis and evaluation methods

3.3.3 Planting area measurement

Auxiliary Datasets Catalogue Sampling

Ground Sampling

Planting Structure Regionalization

Ground Survey Samples

Survey Data PreparationSurvey Method Determination

Survey Result AnalysisSurvey Result

Report

Remote Sensing Datasets

Basic Spatial Framework

Auxiliary Datasets

Ground Survey Samples

Methods of Moderate and Low Resolution

Methods of High-Resolution

Methods of Moderate and High Resolution

Spatial Sampling

Provincial Result

Township Result

County Result

Stratification Mark

Regional

Mesurement

Results

Provincial Result

County Result

Gross Control

Comprehensive Analysis and Evaluation

of Survey Result

Provincial Result

(Report)

County Result (Report)

Township Result

(Reference)

Planting area measurement system of NSRCP

3.3.4 Growth monitoring and yield estimation

The program mainly includes: 1 Regionalization of crop yield estimation

2 Growth monitoring techniques

3 Techniques of crop yield estimation using weather information

4 Techniques of crop yield estimation based on historical statistical data at county-level and remote sensing technology

5 Techniques of crop yield estimation based on real measured yield data and remote sensing technology

6 Techniques of crop yield estimation based on crop growth model and remote sensing technology

3.3.4 Growth monitoring and yield estimation

Geo-spatial Framework

Data Preparation Yield Estimation ModelingYield Estimation Result

AnalysisYield Estimation Result

Submission

Meteorological Yield Estimation

Model

Yield Estimation Based on County Statistical Data

Yield Estimation Based on Ground Measured Data

Crop Growth Simulation Model

Yield Estimation Regionalization

Regional Estimated Yield

Provincial Estimated Yield

County Estimated Yield

Growth Monitoring Model

Other Multi-Source Remote

Sensing Dta

Administrative Boundary of All

Levels

Measurement Data of Sampling

Blocks

Historical Comparison of County Growth

Spatial Distribution of

Provincial Growth

Historical Comparison of

Provincial Growth

Comprehensive Analysis of Growth Monitoring Results

Comprehensive Analysis of All Models' Results

Crop acreage Distribution

Other Thematic Statistical Data

Remote Sensing

Dtasets

Agricultural Statistical

Thematic Data

Basic Geographic

Data

Field Samples

Figure 5 Crops growth monitoring and yield estimation technology system of NSRCP

3.4 Applications of NSRCP

Since the establishment of NSRCP, Beijing Normal University, Chinese Academy of Agricultural Sciences and local coordination institutions, supported by the National Bureau of Statistics, have tested run the system in Jilin, Henan, Jiangsu, Hubei and Beijing, and gradually extended to 13 major grain-producing provinces in China.

Here is an example in Jiangsu province.

3.4.1 Winter wheat acreage measurement in Jiangsu province in 2009

Date : From May 21 until June 20, the National Bureau of Statistics, Beijing Normal University and other research institutes took about 30 days to complete the winter wheat acreage measurement in Jiangsu province in 2009.

Data : present HJ and ALOS, history SPOT as base map, high precision block maps, administrative division maps, traffic information, phenological calendar, planting structure and meteorological data.

Result : With multi-level multi-scale sampling and field survey, winter wheat planted area in Jiangsu province in 2009 is 3225.0 mu, which is calculated out by sampling back stepping.

Date : From May 21 until June 20, the National Bureau of Statistics, Beijing Normal University and other research institutes took about 30 days to complete the winter wheat acreage measurement in Jiangsu province in 2009.

Data : present HJ and ALOS, history SPOT as base map, high precision block maps, administrative division maps, traffic information, phenological calendar, planting structure and meteorological data.

Result : With multi-level multi-scale sampling and field survey, winter wheat planted area in Jiangsu province in 2009 is 3225.0 mu, which is calculated out by sampling back stepping.

3.4.1 Winter wheat acreage measurement in Jiangsu province in 2009

Multi-level multi-scale sampling plots at province level

3.4.1 Winter wheat acreage measurement in Jiangsu province in 2009

County Results

Town Results（ Reference）

Spatial distribution of the winter wheat in Jiangsu province in 2009

3.4.2 Winter wheat yield measurement in Jiangsu province in 2009

Date : From May 15 until June 31, the National Bureau of Statistics and Beijing Normal University took 15 days to complete the winter wheat yield measurement in Jiangsu province in 2009.

Data : MODIS in the growing season of winter wheat, present HJ and ALOS, regionalization map of yield estimation, administrative division maps, real measured yield data, historical statistical data at county-level and meteorological data.

Result : We predict that yield of winter wheat in Jiangsu province in 2009 is 5334.12 kg / ha, which is a little lower than in 2008, decrease degree is about 0.79%.

Date : From May 15 until June 31, the National Bureau of Statistics and Beijing Normal University took 15 days to complete the winter wheat yield measurement in Jiangsu province in 2009.

Data : MODIS in the growing season of winter wheat, present HJ and ALOS, regionalization map of yield estimation, administrative division maps, real measured yield data, historical statistical data at county-level and meteorological data.

Result : We predict that yield of winter wheat in Jiangsu province in 2009 is 5334.12 kg / ha, which is a little lower than in 2008, decrease degree is about 0.79%.

3.4.2 Winter wheat yield measurement in Jiangsu province in 2009

Yield of winter wheat in Jiangsu in 2009

4. Summary and outlook

4.1 Summary

Problems for international major research programs:1.Main Research methods are sampling combined with remote sensing to obtain large-scale crop acreage and yield, which cannot provide results for all levels of government departments to meet the requirement of national statistical offices. 2.there are many problems in these systems, such as indicator, statistical unit classification, survey system, survey accuracy and so on. It is difficult for these systems to enter the national survey system.NSRCP solve these problems to a certain extent:1.Expand current survey indicators comprehensively, provide relatively accurate and objective results, and gradually meet the growing demands of national statistics and the public, under the promise that the timeliness of statistical survey is ensured.2.The system combine the advantages of the country's current survey system and "3S" technology comprehensively.

Problems for international major research programs:1.Main Research methods are sampling combined with remote sensing to obtain large-scale crop acreage and yield, which cannot provide results for all levels of government departments to meet the requirement of national statistical offices. 2.there are many problems in these systems, such as indicator, statistical unit classification, survey system, survey accuracy and so on. It is difficult for these systems to enter the national survey system.NSRCP solve these problems to a certain extent:1.Expand current survey indicators comprehensively, provide relatively accurate and objective results, and gradually meet the growing demands of national statistics and the public, under the promise that the timeliness of statistical survey is ensured.2.The system combine the advantages of the country's current survey system and "3S" technology comprehensively.

Since its inception, NSRCP has been tested run in demonstration provinces such as Jilin, Henan, Jiangsu, Hunan and Beijing.

4.2 Outlook

Through testing run in these provinces, we found that the system still needs major breakthrough at the following two points: (1) Improve the ground-support network: Set up a wireless sensor network in crop measurement area to monitor crop conditions at real time, provide accurate real-time parameters for crop acreage and yield measurement models, which can reduce the cost of field investigation and exclude the interference of man-made factors. (2) Increase the accuracy of crop acreage and yield estimation: Improve remote sensing measurement methods by using multi-source multi-scale remote sensing data and real-time crop growth information provided by wireless sensor network. Improve measurement accuracy of sampling plots to improve the reliability of the result of multi administrative level.

Through testing run in these provinces, we found that the system still needs major breakthrough at the following two points: (1) Improve the ground-support network: Set up a wireless sensor network in crop measurement area to monitor crop conditions at real time, provide accurate real-time parameters for crop acreage and yield measurement models, which can reduce the cost of field investigation and exclude the interference of man-made factors. (2) Increase the accuracy of crop acreage and yield estimation: Improve remote sensing measurement methods by using multi-source multi-scale remote sensing data and real-time crop growth information provided by wireless sensor network. Improve measurement accuracy of sampling plots to improve the reliability of the result of multi administrative level.

Thank you!