Agriculture


What about the effective cropped area and crop production figures? Simply inconsistencies.


Often, during most of the crop season, agricultural areas are principally the soil on which crop will grow later on. From the analysis of optical multi-spectral images, it is a general agreement, that it is not possible to identify on-going field preparations such as ploughing / re-ploughing, sowing, and the earliest stages of plant emergence. Hence, first estimates of the actual cultivated area are not available earlier than crop start developing their plant structure, assuming that weather conditions are favourable to the data acquisition. Nevertheless, the specific sensitivity of Synthetic Aperture Radar (SAR) to important soil properties, such as roughness and moisture content can be exploited. These properties of soil as well as the evolution over time is not casual, as far as agricultural surfaces are concerned. In fact, knowledge of crop calendar, land practices, and precipitation data, multi-temporal SAR data offer valuable information to determine at the earliest stage of the crop season, when and where fields are prepared, and later, the phenological crop status such as flowering, ripening, plant drying and harvesting.


Remote sensing-based information and insurance for crops in emerging economies (RIICE)

Rice and poverty coincide in Asia, home to over 70 percent of the world poor (900 million people) and where almost 90 percent of the world rice is produced and consumed. The Asian population continues to grow rapidly, meanwhile commodity prices are rising and the available arable land area is decreasing. In most of the developing world, rice availability is equated with food security and closely connected to political stability: Rice price increases have caused social unrest in several countries, most recently during the food crisis of 2008.

The objective of this project is to reduce vulnerability of smallholders engaged in rice production by setting up an easy accessible rice information system which opens the way for involved public and private stakeholders to better manage domestic rice production and the risks it is exposed to (development of crop insurance solutions). In the long run rice yields should increase due to better access to information about the actual growth status of observed rice crops and the forecasted yields (as well as about damages and forecasted losses of rice crops), hence leading to a better land management by farmers. Additionally, crop insurance take-up by smallholders facilitates their negotiation position in applying for loans which eventually leads to increased investments in their agricultural business. To reach this, a multiphase project is proposed that will focus on major rice growing areas in selected Asian countries (Bangladesh, Cambodia, India, Indonesia, the Philippines, Thailand, and Vietnam) in the first three years. In the following three years, the activities will be up-scaled to the remaining major rice producing areas of Asia. Objectives for the first phase are:

1. Provision of reliable rice production information in major rice growing areas.
2. Transfer of appropriate know-how and remote sensing technology to national partners.
3. Development of a model aiming at improving production forecast by combining remote sensing, in situ and climatic data.
4. Setting up sustainable crop insurance schemes by developing insurance solutions covering production shortfalls.
5. Provision of crop insurance solutions for at least 5 million rice growing farmers.

The basic idea behind the generation of rice acreage using SAR is the analysis of changes in the acquired data over time. Measurement of temporal changes of SAR response due to the rice plants phenological status - an increase in the SAR backscatter corresponds to a growth in the rice plants - lead to the identification of the areas subject to transplanting / emergence moment and the rice growth. The rice acreage statistics are stored in map format showing the rice extent and, in form of numerical tables, quantifying the dimension of the area at the smallest administrative level - typically village unit - cultivated by rice. These products are linked to district, region, province and country, so that statistics on any of these administrative units can be produced. The figure shows a typical output.

Rice yield prediction is performed by combining remote sensing, in situ, climatic data and an Agro Meteorological Model. Production (t), finally, is simply calculated by combining yield estimation (t/ha) and the acreage (ha) derived from the SAR data.

A public-private partnership consortium is implementing the project of which Swiss Development Cooperation (SDC) will be one of the consortium members. The consortium is composed of sarmap providing the necessary remote sensing technology; IRRI (International Rice Research Institute) is the public partner and will provide a rice crop growth model and work with regional partners to put the system up and running at national levels; AllianzRe Switzerland supported by the Deutsche Gesellschaft fuer Internationale Zusammenarbeit (GIZ) will develop insurance solutions based on the information provided by sarmap and IRRI and pass those solutions on to interested national partners as crop insurance schemes. Besides its financial contribution, SDC role (supported by GIZ) is to institutionally and politically support the partners by facilitating the relations to relevant ministries in targeted countries.


Global Monitoring for Food Security

Existing Food Security Information Services based on Earth Observation generally provide following small scale information:

  • Qualitative vegetation indices primarily used for trend analysis;
  • Water stress;
  • Estimated rainfall;
  • Sporadic ground observations of agricultural variables.

The Global Monitoring for Food Security service contributes to improved food security in Africa, by providing an Earth Observation service based on the integration of low (1 km) to high (10 m) resolution optical and SAR data. The key activities are:

  • Facilitating access to Earth Observation data;
  • Providing an operational service;
  • Validating Earth Observation products;
  • Capacity building.

In Africa, off crop season, crop areas are principally bare soil on which maize will grow later on. After ploughing and planting phase, dependent upon land practices and weather conditions, maize starts developing their plant structure. Months later, after the vegetative and reproduction stage, plants dry before they are harvested. The specific sensitivity of the radar backscatter to soil properties, such as roughness and moisture content, makes possible the detection of these changes already at the earliest stage (e.g. ploughing, sowing, and plant emergence). During the second phase (namely flowering to plant drying stage), the information content of high frequency SAR data and optical data is highly correlated: in the optical case the strong reflectance is due to the high chlorophyll content of the plant, while later to the loss of chlorophyll (drying stage). In the SAR case, the plant humidity and mainly the volume scattering are the key factors determining the high reflectivity. The lower reflectivity during the plant drying is exclusively caused by the loss of wetness in the plant. The figure shows a typical set of output generated using high resolution SAR data.

The color composite on the left illustrates a multi-temporal data set based on 120 ENVISAT ASAR AP images and 70 ALOS PALSAR FBS scenes data covering the whole Malawi (100,000 sqkm, 15m resolution). The image on the right shows an interferometric color composite based on ALOS PALSAR Fine Beam Single data (70 image pairs). The enlargements highlight the extensive information included in this type of multi-temporal multi-source data set, which allows the generation of products such as crop map, main land cover/change classes, and Digital Elevation Model. All processing has been performed starting from SLC data.


In the past years, with the launch of very high resolution SAR systems - such as the Cosmo-SkyMed constellation - the estimation of the cultivated area could be additionally enhanced by integrating Cosmo-SkyMed (3m), ENVISAT ASAR (15m), and ALOS PALSAR-1 (15m) data. The method consists in the generation of three independent and complementary products, which, in turn, they are fused, enabling the derivation of the cultivated area one. Each intermediate product has a clear meaning within agriculture and food security: i) the potential crop extent prior to the crop season; ii) the potential area at start of the crop season; iii) the crop growth extent during the rainfed crop season.

Malawi-Lilongwe - Intermediate products (top) and Cultivated Area product (bottom). Top left - Potential Crop Extent prior to the start of the rainfed crop season is derived from ALOS PALSAR-1 FBD (15m) multi-annual intensity acquired during the dry season. Top center - Potential Cultivated Area at Start of Season is derived from one-day interferometric Cosmo-SkyMed data (3m) acquired during the fields preparation period. Top right - Crop Growth Extent is derived from multi-temporal ENVISAT ASAR AP/IM intensity data (15m) acquired during the crop season. The three maps correspond to the three selected ENVISAT ASAR acquisition modes IS2, IS3, IS4.



Selected publications

S.P. Kam, F. Holecz, E. van Valkengoed, M. Barbieri, C.B. Casiwan, S.L. Asilo, L. A. Santos, R.G. Manalili, W.B. Collado, S.A. Adriano, and A. Maunahan, The makings of an internet-based rice information system: Piloting in the Philippines, First Symposium on Geoinformatics, Philippines, 2004.
F. Holecz, M. Barbieri, A. Cantone, P. Pasquali, and S. Monaco, Synergetic Use of multi-temporal ALOS PALSAR and ENVISAT ASAR data for Topographic/Land Cover Mapping and Monitoring at National Scale in Africa, IGARSS Symposium, Cape Town, 2009.
F. Holecz, F. Collivignarelli, M. Barbieri, L. Copa, and S. Monaco, Estimation of cultivated area in small plot agriculture in Africa for food security purposes, First Announcement of Opportunity COSMO-SkyMed Workshop, Rome, 2012.