Assessing climate risks to soybean yields: Effects of extreme air and land surface temperatures across critical growth stages

In the context of climate change, with an expected increase in heat waves, longer warm seasons and shorter cold seasons, understanding how the increased frequency of extreme temperature events affects crop yields is essential to mitigate their negative impacts. In this paper, we investigate how climatic temperature extremes have affected soybean yields in a 17-year series in the Argentine Pampean region, through the estimation of indices of temperature extremes, including maximum and minimum temperatures, and a representative index of diurnal temperature range. The indices were calculated from air temperature measured at weather stations, air temperature derived from reanalysis models and land surface temperature (LST) obtained from satellites in areas compatible with soybean cultivation. The frequency of high maximum temperatures was negatively associated with yields (regional average r = -0.65 for LST, up to −0.8 in several locations). The correlations were stronger when using extreme temperature indices based on the 90th percentile, compared to those derived from fixed thresholds. In contrast, low maximum temperatures were positively correlated with soybean production. The relationships were stronger when considering land surface temperature instead of air temperature. In most cases, minimum temperatures did not have a significant effect on soybean yields in the areas studied. Diurnal temperature range showed predominantly negative relationships with yields, mainly at the start of the critical period of soybean. In the context of climate change, our findings highlight the value of integrating satellite-based thermal data into monitoring frameworks to enhance resilience, reduce risks, and support the long-term sustainability of agricultural systems. The data from meteorological stations used in this work were provided by the National Meteorological Service ( https://www.smn.gob.ar/ ) and the National Institute of Agricultural Technology ( https://siga.inta.gob.ar/#/ ). They are available for research purposes upon request. ERA5 data were downloaded from ERA5 reanalysis data were downloaded from the KNMI Climate Explorer and European Centre for Medium-Range Weather Forecasts ( https://climexp.knmi.nl/start.cgi ). Surface temperature data were downloaded from Earth Science Data Systems ( https://search.earthdata.nasa.gov/search ). • Extreme temperature indices were calculated from air and land temperatures. • Soybean yields were negatively correlated with maximum temperature indices. • Percentile-based indices showed stronger correlations than fixed-threshold ones. • Thermal amplitude in the early critical period is well correlated with soybean yields. • Satellite-based land surface temperature indices had the strongest yield correlation.