Estimating drought-induced crop yield losses in near-real time

Authors: MEITNER, J - Global Change Research Institute; BALEK, J - Czech Globe - Global Change Research Institute; BLÁHOVÁ, M - Mendel University; SEMERÁDOVÁ, D - Global Change Research Institute; HLAVINKA, P - Mendel University; LUKAS, V - Global Change Research Institute; JURECKA, FRANTISEK - Mendel University; ŽALUD, Z - Mendel University; KLEM, K - Mendel University; Anderson, Martha; DORIGO, W - Vienna University Of Technology; FISCHER, M - Czech Globe - Global Change Research Institute; TRNKA, M - Mendel University

Submitted to: Agronomy
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/17/2023
Publication Date: 6/21/2023
Citation: Meitner, J., Balek, J., Bláhová, M., Semerádová, D., Hlavinka, P., Lukas, V., Jurecka, F., Žalud, Z., Klem, K., Anderson, M.C., Dorigo, W.A., Fischer, M., Trnka, M. 2023. Estimating drought-induced crop yield losses in near-real time. Agronomy. 13(7):1669. https://doi.org/10.3390/agronomy13071669.
DOI: https://doi.org/10.3390/agronomy13071669

Interpretive Summary: Federal drought-induced yield loss compensation programs require reliable methods for ascertaining extent and severity of drought impacts on crop production. To date, drought compensation programs in the Czech Republic have been based on yield reporting data acquired at low spatial resolution (large district scales) or with excessive time lags. This paper describes the development of an improved crop loss monitoring protocol for the Czech Republic, using a combination of modeled and remotely sensed indicators of soil moisture and crop stress within a machine learning model. The study also integrates near-real-time feedback from a network of reporters and from crop status surveys. The system is demonstrated for the growing seasons of 2017 and 2018, which were severely drought impacted in the Czech Republic. Reasonable estimates of yield loss for 17 different crops were obtained at the cadastral level, with 30% loss thresholds used to identify areas qualifying for payments. Study results indicate that it is very appropriate to use such a yield loss model as a first criterion for allocating compensation.

Technical Abstract: In the Czech Republic, soil moisture content during the growing season has been decreasing over the past six decades, and drought events have become significantly more frequent. In 2003, 2015, 2018 and 2019, drought affected almost the entire country, with droughts in 2000, 2004, 2007, 2012, 2014 and 2017 having smaller extents but still severe intensities in some regions. The current methods of visiting each individual cadastral area (there are approximately 13,000 of them) to allocate grants for the crop yield losses caused by drought or aggregating the losses to district areas (approximately 1,000 km2) are both inappropriate. For this reason, the development of a method combining ground surveys and remotely sensed or model data for determining crop yield losses is necessary. Such a method is presented in the current study. First, the study shows that it is possible to estimate crop yield losses at the cadastral area level in the Czech Republic and attribute those losses to drought. This can be done by combining remotely sensed vegetation, water stress and soil moisture conditions with modeled soil moisture anomalies, which are coupled with near-real-time feedback from a network of reporters and with individual crop status surveys. The data are then combined through artificial neural network models, and a scheme that enables the rapid estimation of crop yield losses as a consequence of real droughts is developed.