Yield loss estimation v2
- 1. CROP LOSS ESTIMATION SONU AGRAWAL MANAGING DIRECTOR sonu.agrawal@weather-risk.com SEPTEMBER 2014
- 2. ABOUT US YIELD ESTIMATION IN INDIA To provide all in the world security against climate change as the world’s No. 1 climate risk management company using data, technology and financial services Founded 2004 Headquarters India Team 101 People Footprint Pan India & Global Asia Philippines Cambodia Bangladesh Sri Lanka Africa Tanzania Rwanda Zambia Mozambique Existing Presence Building Presence SIDBI, IIT Kanpur Ford Foundation, ILO HNIs VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 3. Agricultural Technology Innovations • Disease Forecasting • Agricultural Goods Tracking • Smart Irrigation Systems 11 years of sterling track record in Crop Insurance Founded NDVI Index solutions CCE surveillance solutions 2004 2006 2008 2010 2012 2013 & 2014 Insurance for Coffee growers Varsha Beema for Rajasthan Together Convinced GoI to subsidize weather insurance with Rs. 1000 Cr. AWS Grid ~ 100 Stations Data services for Bengal, Rajasthan, Uttar Pradesh 2005 2007 2009 2011 Cloud Cover for Bayer Seeds ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 4. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING Requirements Challenges Mid Season yield assessment (check if yield is likely to be less than 50%) No uniform established methodology If sown area is < 25% of the normal; Easier to do; but problems in flood CCEs at GP level 250 audits per district in 15 days???
- 5. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE Android based smart phone application MODELING
- 6. In-Season Crop Damage/Loss Yield Estimation Dividing crop period into different vegetative and reproductive stages. Vegetative stages – counted as consecutive unfolded leaves, until the reproductive parts are visible on the plants. Reproductive stages – as soon as the flowers/tuber/ear head are visible until all the kernels/seed/tuber are physiologically mature. Crop damage based on parts of the crop which is damaged. Crop Stand damage – Count or percentage of crop stand area with no living axils/buds. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 7. In-Season Crop Damage/Loss Yield Estimation Crop damage based on parts of the crop which is damaged. Crop stem damage – Count and percentage of crop stem snapped off with physiologically unable to produce yield or inactive. Branch damage – Position and percentage of branches snapped off or damaged. Leaf damage - Count and percentage of leaves are snapped off, shredded, de-colourized and physiologically inactive that wilts and dies. Ear/Pod/Head/Boll damage – Count and percentage of yield part knocked off/chaffed/shriveled/ broken or disease/pest infected. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 8. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING In-Season Crop Damage/Loss Yield Estimation Fruit damage Count and percentage of fruits/ tree knocked off/ malformed/ disease/pest infected and quality degraded. Crop Yield estimation before harvesting period. Locating representative sample area. Determining the plant stand and row width. Determining the plants (or ear/ fruit/ pod) sample population / 100 sq mt. Filling observation report. Estimating the yield based on observations. Forecast Yield (Y) = F(seed weight, plants, row width) Y = …… t/ha Yield Loss = Normal Yield – Forecast yield UAV & SATELLITE MODELING
- 9. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 10. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 11. ABOUT US YIELD ESTIMATION IN INDIA UAV Images from 100 meters VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING Object- based hierarchical image analysis to classify imagery of plots measured concurrently on the ground using standard rangeland monitoring procedures.
- 12. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV Images using regular camera from 250 meters UAV & SATELLITE MODELING Objects are further classified into vegetative groups and to species level by Rule Based Classification with well defined thresholds and Near Neighbor Classification Algorithm, feasible for few crops.
- 13. VIDEOGRAPHY AND CLOUD SOURCING Paddy ABOUT US YIELD ESTIMATION IN INDIA UAV Images using spectral camera from 250 meters UAV & SATELLITE MODELING
- 14. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 15. Methodology Monitor Yields through Satellite images (LISS4, LANDSA, SAR). We use LANDSAT images of resolution 30m*30m. In case, of more detailed analysis, will use LISS4 images of 5m*5m resolution. Where visibility is affected due to clouds, Microwave SAR data can be used. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 16. Methodology Stasny-Goel method (Bayesian method) Bayesian yield estimation algorithm with a simple spatial component based on the crop yields (close geographic proximity tend to be more similar than those further apart). Griffith method (AR model) Box-Cox and Box-Tidwell transformations are employed in conjunction with an autoregressive specification so as to optimize agreement with model assumptions. Standard ratio estimation Multi-Phase stratified sampling is used to generate ratio estimates that are weighted by the sampling rate. Econometric methodology Use of Econometric principles and model building by considering endogenous and exogenous variables such as prices of both product and inputs, farmer planting decision and consumer preferences, etc.. Agro-Met methodology These model are crop growth stimulation model which is a function of complex interaction of different physiological processes with the environment, biotic and a biotic factors. Geographic Information System and Remote Sensing Methodology These model use RS and GIS information for quick assessment based on multispectral, large area. Best for taking critical decision on procurement, transportation, storage and trade. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 17. Frame work for Integrated Crop Yield Forecast Crop Survey, AWS and Satellite Data Growth Monitoring - Crop phenology, Vegetative Index (NDVI) Agro Meteorological Data – Precipitation, PAR, Water Holding Capacity, GDD Input Yield Per unit Area (YPA) Empirical Method YPA = f(xi) Xi = Meteorological indices, Drought Index, Vegetative indices Crop yield estimate Yield = f(NDVI , Rain Index, GDD, Ancillary data) ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING
- 18. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE Integrated Yield Estimation Model Crop yield can be estimated by adopting advance technologies such as remote sensing imagery (RS), Geographical Information system (GIS), etc. and appropriate methodologies such as Multivariate regression. Crop models with other important inputs from weather data, land based observations and economic parameters that influence the farmer’s decision on cultivating particular crop. A general integrated yield estimation model for estimating crop yield. Y = f (RD, RFs, W, S, Pt, I, G) Y = Yield of the crop. RD = Remote sensing imagery data (NDVI, SAR, IRS-WiFS, etc.) RFs = Rainfall received during sowing and Vegetative stage. W = other important weather parameters. Pt = Previous year yield of the crop. S = Soil type and its Parameters I = Irrigated area availability. G = Ground truth data by using CCE (crop cutting experiment) approach. MODELING
- 19. ABOUT US YIELD ESTIMATION IN INDIA VIDEOGRAPHY AND CLOUD SOURCING UAV & SATELLITE MODELING Sl.No. Parameters Source 1. Remote Sensing Imagery Landstat, NRRS, UAV’s and Aerial vehicles 2. Weather Data Ingen AWS, IMD, TRMM, GFS, NASA 3. Previous year’s Data State Govt. DES 4. Fertilizer and Input Details State Agriculture Department 5. Soil Details NBSS & LUP 6. Irrigation Details Central Ground Water Board 7. Ground truth Data Weather Risk (Field Survey data)
- 20. 20 The sky is not the limit SONU AGARWAL MANAGING DIRECTOR sonu.agrawal@weather-risk.com SEPTEMBER 2014