Efficacy of Satellite Data use in PMFBY — An Analysis
By Prateep Basu, Co-Founder, and CEO, SatSure
The Government has notified key changes to the Pradhan Mantri Fasal Bima Yojana (PMFBY) in September 2018, including several changes to the operational guidelines aimed at streamlining the whole process. Most importantly, the Government has mandated the use of technology including satellite-based remote sensing for monitoring of Crop Health/Crop Cutting Experiments (CCEs) in coordination with concerned States.
Contained in the provision, the insurance companies, and concerned states shall use as satellite and UAV remote sensing, for various applications such as Crop Area Estimation and Yield Disputes and also promote the use of remote sensing and other related technology for CCE Planning, Yield Estimation, Loss Assessment, Assessment of Prevented Sowing and Clustering of Districts. The states are mandated by the provisions of the revised guidelines to provide Insurance companies with the prior approval of the agency from which such data can be procured. This is required for better monitoring and ground-truthing.
In addition, there is a slew of penalties for States and Insurance Companies (IC) for the delay in settlement of insurance claims under the scheme. This crucial provision will see the farmers being paid 12% interest by insurance companies for the delay in settlement claims beyond two months of the prescribed cut-off date. State Governments will have to pay 12% interest for the delay in the release of the State share of subsidy beyond three months of prescribed cut-off date submission of requisition by insurance companies. The new operational guidelines come at the onset of the rabi season.
The new operational guidelines also detail a Standard Operating Procedure (SOP) for evaluation of insurance companies and remove them from the scheme if found ineffective in providing services. The Government has also decided to include perennial horticultural crops under the ambit of PMFBY on a pilot basis. The scheme, as per the new operational guidelines provides add-on coverage for crop loss due to the attack of wild animals, which will be implemented on a pilot basis. Aadhaar number will be mandatorily captured to avoid duplication of beneficiaries.
In order to ensure that more non-loanee farmers are insured under the scheme, apart from various awareness activities being scheduled, the insurance companies are given a target of enrolling 10% more non-loanee farmers than the previous corresponding season. The insurance companies will have to mandatorily spend 0.5% of gross premium per company per season for publicity and awareness of the scheme.
The new operational guidelines address the current challenges faced while implementing the scheme by putting forth effective solutions. The much-demanded rationalization of the premium release process has been incorporated in the new guidelines. As per this, the insurance companies need not provide any projections for the advance subsidy. Release of upfront premium subsidy will be made at the beginning of the season based on 50% of 80% of the total share of subsidy of the corresponding season of the previous year as GOI/State subsidy. The balance premium will be paid as a second installment based on the specifically approved business statistics on the portal for settlement of claims.
The final installment will be paid after the reconciliation of entire coverage data on the portal based on final business statistics. This will reduce the delay in settling the claims of farmers.
Below is a non-exhaustive list of Satellite-based Remote Sensing use in PMFBY as mentioned in the revised guidelines.
Rationalizing CCEs
With the availability of a number of satellites with high-resolution imaging capability orbiting the Earth, there is a great improvement in satellite remote sensing-based products. It has been reasonably proven that satellite-based vegetation indices (such as Normalized Difference Vegetation Index, NDVI and Normalized Difference Wetness Index, NDWI) can help in demarcating the cropped areas into clusters on the basis of crop health. This feature can be successfully used to target the CCEs within the Insurance Unit (IU). In other words, satellite imagery can help in ‘smart sampling’ (stratified sampling) of CCEs. This will help in optimizing the number of CCEs, to make them representative of different crop conditions. This is expected to reduce the total need for CCEs by about 30–40% while maintaining similar accuracy.
It will also give a more representative yield of the IU, as it will consider all crop pixels (fields) within the IU and not just the location of 4 CCEs. States can adopt this technique (of using satellite-based remote sensing data for planning of Crop Cutting Experiments) in generating yield estimates. For using satellite data for smart sampling, there is a need to generate a specific crop map. The MNCFC under the FASAL project of the DAC&FW prepares crop maps for various crops.
These crop maps can be used for CCE planning. Vegetation Indices (NDVI and NWDI) need to be computed for the cropped area. On the basis of the Vegetation Index, the crop area can be categorized into poor, medium, good, and very good crop health strata. Within each stratum, CCE points should be selected randomly.
Additionally, ground-truthing within these IUs can be done in order to develop a crop yield model to ascertain yield estimation of the crop and simultaneously can act as a reconciliation/verification tool of actual CCEs conducted on the field. This can be done in consultation with MNCFC, NRSC (ISRO), SAC (ISRO), SRSC, and IASRI. These Departments have also been piloting such studies, in this regard, for optimization/reduction of crop cutting experiments using technology.
Removing Area Discrepancy In Coverage
Area Insurance in some states and districts sometimes exceeds the area sown. To correct this discrepancy use of remote sensing technology and satellite imagery, digitization of land records needs to be done.
Direct Yield Estimation
For addressing the issue of reliability of CCEs in terms of their accuracy, representativeness, and timeliness, innovative technologies such as satellite remote sensing, drone, modeling, AWS/ARG, real-time transmission of data, etc. should be utilized.
This will ensure an accurate assessment of yield and timely payment of claims to farmers. Various studies carried out by national and international organizations, including MNCFC, NRSC, SAC, CCAFS, IRRI, IFPRI, World Bank, etc. have shown that the use of satellite, weather, soil, and crop data, along with images/video capture of crop growth at various stages and accurate sample CCE data collection can improve the yield data quality/ timeliness and support timely claim processing and payments.
States, with the support of national centers, need to carry out an adequate number of pilot studies for improved yield estimation using technology, mentioned correlation is observed between and yield estimated through CCEs, States and Insurance Companies use these technologies in estimating the crop yields at IU level, subject to the satisfaction of both States and Insurance Companies about the accuracy of the yield estimates, to service the claims.
Improving Yield-data Quality and Timeliness
The normal CCE process being followed by the State for estimating yield is lacking in reliability, accuracy, and speed, which affects claim settlement. There is a need for real-time, good quality, and reliable actual yield data for which mandatory use of smartphones/handheld devices has to be done for capturing images, location of the CCE, and for online transmission of data on the National Crop Insurance Portal through CCEs Agri-app. RST using satellite and drones, weather data, models, etc. may also be used for the purpose of increasing accuracy and speed of yield estimation through CCEs.
The cost of using technology etc. for the conduct of CCE process especially purchase of smartphones/ handheld devices and use of technology (RST, Drone, etc.) will be shared between Central Govt. and State Govts/ U.Ts on a 50:50 basis, wherever necessary, subject to a cap on total funds to be made available by Central Govt. for this purpose based on approximate cost of procuring handheld devices/smartphones and other related costs (RST, Drone, etc).
Proxy Indicators for Mid-Season Adversity
Indicators to be used for loss intimation could be rainfall data, other weather data, satellite imagery, drought assessment reports of MNCFC, and crop condition reports by district level/ State Govt. officials, supported by media reports and field photographs.
Risk Clustering
Clustering of Risk as per 2016 Guidelines of PMFBY, for the purpose of clustering/clubbing of districts and determination of L1 bidder, the risk is analyzed based on long term data of yield by (a) computing average burn cost (percent difference between actual yield and threshold yield) and b) computing the level of variability in long term yield. Since, availability of high-quality long-term yield data is difficult, especially at the lower administrative levels, other methods (including the use of satellite data) can be tried for risk assessment.
Satellite data, of moderate resolution (e.g. Resource- sat AWiFS, Terra/Aqua MODIS) are available, for the long term, i.e. at least around 15 years. The long-term Vegetation Indices, which are indicators of crop health, derived from these satellites can be used to assess the year-to-year variations and thereby understanding the risk potential of an area. It can be combined with many other satellite-derived products, such as flood maps, drought assessments, and vulnerability and long-term weather data to carry out the risk analysis.
Data Integration to the National Portal
Integration of RST/Satellite Data for handling Data Discrepancy/dispute resolution: New Age Technologies like Remote Sensing Technology is a promising step in bringing in procedures and systems of approach which is more reliable, accurate, and fast in resolving errors/concerns of stakeholders and provide a progressive and scientific solution which the traditional/existing procedures and practices are unable to provide. This will help in the reduction in time required for the collection and collation of different data sets and reports pertaining to crop health, productivity, sowing, and harvesting activities.
The following data sources may be used for validation of on account claims and claims for prevented sowing:
• Satellite/UAV Remote Sensing Data
• AWS/ARG Data
• MNCFC Report/Study on drought assessment.
Analysis of Vegetation Index
Remote Sensing Data is the central tool for calculating the vegetation index. For the purposes of PMFBY, the indexes used are Normalized Difference Vegetation Index (NDVI), Normalized Difference Wetness Index (NDWI), Vegetation Condition Index (VCI), Temperature Condition Index (TCI), Vegetation Health Index (VHI), Enhanced Vegetation Index (EVI), etc. To derive these indices, satellite data of appropriate resolution should be used.
Types of Spatial Resolution to be used for vegetation index derivation for different levels of analysis Level of Analysis are the following.
- Village level 5–10 m or better
2. Block/Tehsil level: 10–30 m
3. District level: 50–100 m
Crop Growth Modelling
Yield loss estimates can be made using crop simulation models such as DSSAT/ InfoCrop, etc. A remote sensing-based semi-physical modeling approach can also be used for crop growth analysis. However, care should be taken to use well-calibrated and validated models and also models should be run in spatial at higher resolution, at least 5 km.
Identification of Outliers
All these above analyses can be used to check whether there was any reason for yield deviation as presented in the CCE data. Then a yield proxy map may be prepared. The Yield proxy map can be derived from remote sensing vegetation indices (single or combination of indices), crop simulation model output, or an integration of various parameters, which are related to crop yield, such as soil, weather (gridded), satellite-based products, etc.
Whatever, yield proxies are to be used, it is the responsibility of the organization to record documentary evidence (from their or other’s published work) that the yield proxy is related to the particular crop’s yield. Then the IU level yields need to be overlaid on the yield proxy map. Both yield proxy and CCE yield can be divided into 4–5 categories (e.g. Very good, Good, Medium, Poor, Very Poor). Wherever there is a large mismatch between yield proxy and the CCE yield (more than 2 levels), the CCE yield for that IU can be considered, as outliers.
(The article was originally published in The SatSure Newsletter. Please click the below link to access the other editions of the newsletter)
