Drought is having a devastating impact on crops. Total precipitation, biomass in a pre-drought year and extreme precipitation event are considered as main factors of biomass (Integrated Enhanced Vegetation Index: iEVI) in a drought year (DY). In terms of vulnerability, it is very important not only to assess the decrease in biomass but also its changes between pre-drought year (PDY) and post-drought year (PoDY). Previous field studies suggested that plant species richness (PSR) is positively related with recovery and resistance. Resistance, ability to withstand the perturbation, is calculated as the difference between DY and PDY biomass. Recovery, ability to compensate for the loss of biomass, is calculated as the difference between PoDY and DY biomass. In addition, according to previous studies, resistance, recovery and resilience seemed to depend on land use history and soil types. Resilience, ability to return to its original state following the perturbation, is calculated as the difference between PoDY and PDY biomass. These ecological factors, however, have not been considered in the biomass and the vulnerability of drought event in a large scale study. The assessment of drought impact in large scale is also critical for appropriate representation of extreme dry conditions in a macro-scale model such as general circulation model. Our main objective is to show the effects of PSR, land use history and soil type on EVI. Dataset include (1) EVI from MODIS, (2) monthly precipitation data from GPCC, (3) PSR data from Ellis et al. (2012), (4) soil types from soil qualities for crop production of FAO (2008), (5) crop land and grassland history for 6000 years from HYDE3.1 and (6) percentage of land irrigated area from FAO (2013). The spatial resolution is 5 minutes. “Drought year” in our research is that annual precipitation is less than 2.5% quantile of cumulative distribution function of annual precipitation from 1901-2014 of GPCC. To consider spatial variability, a Bayesian approach is used with the EVI-based resilience as objective variable and dataset 2, 3, 4, 5 and 6 as explanatory variables. We chose the explanatory variables by Bayesian information criterion. Contribution of PSR and soil types for resilience were small. Whereas grassland history contributed to resilience next to precipitation and irrigation.
各種検索
サポート
ヘルプ