Statistical Analysis for Satellite Index-Based Insurance todefine Damaged Pasture Thresholds

Abstract. Vegetation indices based on satellite images, such as Normalized Difference Vegetation Index (NDVI), have been used in countries like USA, Canada and Spain for damaged pasture and forage insurance for the last years. This type of agricultural insurance is called satellite index-based insurance (SIBI). In SIBI, the occurrence of damage is defined through NDVI thresholds mainly based on statistics derived from normal distributions. In this work a pasture area at the north of Community of Madrid (Spain) has been delimited by means of MODIS images. A statistical analysis of NDVI histograms was applied to seek for the best statistical distribution using maximum likelihood method. The results show that the normal distribution (NORMAL) is not the optimal representation and the General Extreme Value (GEV) distribution presents a better fit through the year. A comparison between NORMAL and GEV are showed respect to the probability under a NDVI threshold value along the year. This suggests that a priori distribution should not be selected and a percentile methodology should be used to define a NDVI damage threshold rather than the average and standard deviation, typically of normal distributions.