Copulas modelling of maize yield losses – drought compound events using the multiple remote sensing indices over the Danube River Basin

• Published in: Agricultural water management 2023-04, Vol.280, p.108217, Article 108217
• Main Authors: Potopová, V., Trifan, T., Trnka, M., De Michele, C., Semerádová, D., Fischer, M., Meitner, J., Musiolková, M., Muntean, N., Clothier, B.
• Format: Article
• Language: English
• Subjects: Coupling drought–maize yield losses mechanism; Enhanced Vegetation Index; Evaporative Stress Index; Food security; Four-variate C-vine copula; Peak correlation timing; Relative Soil Water Availability
• ISSN: 0378-3774; 1873-2283
• DOI: 10.1016/j.agwat.2023.108217

Danube countries have witnessed numerous waves of drought events, causing significant agro-economic loss, but three consecutive dry years amplified the debate on how to deal with future drought risk. The European drought of 2022 has shown how important it is to look at food security from an environmental droughts risk assessment approach. The coupling drought–yield losses framework derives from the understanding that all land systems are connected through coupled human and natural systems, and these social, ecological, and agro-economic impacts are the result. Maize is considered a commodity and a staple food in Europe with the largest market share in global maize exports. Drought–heat stress, war and subsequent limitations on Ukrainian trade have created a shortage of maize supply in 2022 over Europe. This study focused on eighteen countries where maize production becomes highly susceptible to drought in the Danube River Basin (DRB; Austria, Bosnia and Herzegovina, Bulgaria, Croatia, the Czech Republic, Hungary, Montenegro, Romania, Serbia, Slovakia, and Slovenia). To understand the coupling drought–yield losses mechanism, time series of maize yield datasets and multiple remote sensing indices were used on arable lands for 278 districts at a high spatial resolution. The main objective of this study was to determine which regions respond to the changes in the rate of evapotranspiration and soil moisture and in which period and how much maize production is affected. The time series of the two-band enhanced vegetation index (EVI2), the evaporative stress index (ESI), and the relative water availability (AWR) were calculated. The spatial evolution of the ESI for 4-week and 12-week time windows, EVI2, and relative soil saturation at the topsoil and rootzone layers demonstrate the progress of agricultural drought under varying agroclimatic conditions and its impacts on maize yields. Our study adopted a novel mechanism-based risk assessment approach using a four-variate C-vine copula in the perspective of modelling yield losses. To assess how much maize production can be limited by drought stress, the weekly dynamics of the strength of bivariate linkage of eight compound modes were provided. The return periods of drought–yield losses signatures in the study region were less than 4 years. The highest chances (once every 2.50–2.86 years) of the occurrence of drought–yield losses signature occurred in Romania, Bulgaria, Slovakia, Bosnia and Herzegovina. The availab