Spot Image JP - Satellites and agricultural statistics

Home

Home > Geoinformation > Agriculture > Satellites and agricultural statistics

Satellites and agricultural statistics

Agricultural statistics provide a yearly picture of farming practices and production.

They help to gauge the impacts of climatic and economic changes, guide strategic choices and target subsidies. Their role is to improve decision-making and market positioning.

Need

Agricultural statistics must provide decision-makers with detailed national and regional figures. They also have a fourfold purpose:

indicate utilized agricultural areas (UAAs) and non-agricultural land
estimate national distribution of cropped area
forecast potential yield for each type of crop
highlight changes from one year to the next and identify trends.

Spot Image solution

Spot Image designs and operates satellite imaging solutions for monitoring of crop growth dynamics. Its domain experts work with an international partner network to develop applied agricultural statistics solutions.

CASE STUDY - Satellites and agricultural statistics in South Africa

A changing agricultural landscape

South Africa has experienced major political reforms since 1994 and these have also influenced agriculture. Today, a large number of farming homesteads operate parallel to a thriving agri-business sector with access to world markets. Types of farmlands are dictated by varying rainfall patterns, so South Africa needs precise agricultural statistics covering its entire territory.

Limits of initial statistical surveys

In 2002, the South African Ministry of Agriculture requested bids to provide crop estimates for South Africa. In reply to this tender, GeoTerraImage, SiQ and ARC formed the National Crop Statistics Consortium (NCSC). Based on area frame to guide interviews with farmers, the initial statistics and results proved successful.

However, over a period of 5 years, the farmers became reluctant to participate, and the response rate dropped. Also, due to the, cost of the logistics of covering vast areas, alternatives were required. The NCSC started to investigate different options for approaching statistical surveys.





varying rainfall patterns

New methodology

GeoTerraImage used its experience and expertise in remote sensing and geospatial technologies to process and prepare hundreds of SPOT images for digitizing field boundaries as well as crop type classification, helping the consortium to develop an applied agricultural statistics solution using SPOT satellite imagery.

It began by compiling a baseline stratification of fields, encompassing all of South Africa. Seasonal satellite surveys in combination with aerial statistical surveys then worked from this stratification to establish annual agricultural statistics.

A baseline covering South Africa’s 9 provinces

Fields are digitized from 2.5-metre SPOT 5 orthoimages. This digitized field database covers 13 million hectares nationwide.





Fields are digitized

Stratification based on the field boundaries captured from the SPOT 5 images improves sampling efficiency, thereby optimizing the organization and cost of field and aerial surveys.





Stratification

Radiometric samples are collected by aerial statistical surveys for each type of crop from target fields during different seasons to determine crop profiles.





Radiometric samples

Image acquisition planning is adjusted in accordance with crop growth stages.





planning

Multi-seasonal satellite surveys

Survey campaigns are conducted at different stages of crop growth throughout the year. Satellite imagery is orthorectified and processed by GeoTerraImage.





Survey campaigns

Measurements are derived from imagery using NDVI, Tasseled Cap and vegetation indices as well as PCA. These measurements in combination with spectral bands are then classified by GeoTerraImage.





Measurements

Classification is applied to the field database. Statistical analysis—acreages, zoning, distribution, potential yields, etc.—can now begin.





Survey campaigns

Results

Agricultural statistics can now be compared to results for 2007 over South Africa’s entire territory, per province and district. The example here shows changes in areas of maize and soya bean between 2007 and 2008, per district in the province of Free State.

Benefits

Fanie FERREIRA, GeoTerraImage

Food security is of increasing concern on a global scale and the requirement for timely and accurate information about crop availability and supply is more critical than ever before.

The SPOT 5 satellite has several advantages which makes it very useful to support information extraction for Agricultural applications such as crop area estimation.





Fanie FERREIRA - GeoTerraImage (http://www.geoterraimage.com)

Firstly the combination of the two sensors on Spot5 that simultaneously record both in 10m multispectral and 2.5m panchromatic is well suited for agricultural applications. The multispectral sensor at 10m, with near-infrared and mid-infrared bands, allows detailed analysis of a landscape to locate agricultural activity and identify cultivated fields, whilst the 2.5m panchromatic imagery provides sufficient detail to support precise delineation of the perimeter of agricultural fields.

Secondly the large footprint of the SPOT 5 sensors (60km x 60km) has allowed the recording of a complete coverage of cloud free Spot5 imagery for South Africa in three consecutive years (2006, 2007 & 2008). Based on this multi year coverage at very high resolution, the National Crop Statistics Consortium (NSCS) captured and digitized almost 800 000 parcels of agricultural fields covering 13 million hectares across the entire country.

This detailed inventory of agricultural field location and extent provides a stratification framework with several opportunities for annual crop statistics reporting. The primary use of the field framework is for sample selection in preparation of the annual grain crop surveys to determine crop specific area estimates per province in South Africa within the growing season. This survey, known as the Producer Independent Crop Estimate System (PICES), is performed during February/Match for summer grain crops, and during August for winter grain crops. A secondary procedure is the end of season crop type image based classification, to generate a full coverage field level crop type classification, for district level analysis and reporting.

These crop estimates provided to the South African Crop Estimates Committee (CEC) in support of their monthly release, has improved crop estimates figures and subsequently contributed to price stability.

Top