We published a paper on the “Impacts of compound hot–dry extremes on US soybean yields” in Earth System Dynamics. Our results show that soybean yields are most negatively impacted by the combination of high temperature (T) and low soil moisture (SM) during summer. Figure below shows observed (dots) and modelled (contour-lines) yield values. Note the increasing sensitivity of crop yields to high T during low SM conditions. This depicts the synergistic nature of compound weather and associated impacts.
To explore how these hot-dry compound events come to be, we calculate composites of interannual correlations between SM, T and actual evapotranspiration (ET) pairs for a given month of the year, repeated over the various calendar months (see figure below). This allows to compare the land-atmosphere feedbacks during hot-dry summer years vs normal climatological summer years.
Hot-dry events are characterized by stronger negative spring coupling between ET & SM and between SM and T, leading to fast SM depletion in spring and a reversal in the land surface cooling mechanism over summer, prompting important soybean yield impacts. Climate models project substantially warmer summers for the US, although uncertainty remains as to whether this will be accompanied by drier conditions. This highlights a critical element to explore in future studies focused on US crop production risk under climate change.
This research started with the main aim to identify when during the growing season are soybean yields most sensitive to weather and climate variability. A first step was to iteratively test out monthly predictors and select a few that maximized the signal of the statistical model we are trying to build. This selection procedure highlighted the summer period and in particular August and September hot-dry conditions as key variables that affect soy yield variability. An interesting question that came up during the revision phase is whether these conditions are setting up earlier in the season. By investigating this further, we show that summer hot-dry conditions are often coinciding with springs that can be characterised by above average couplings between SM and T. Such feature of the climate system leads to faster soil moisture depletion, a mechanism that reinforces important land-atmosphere feedbacks in the summer further fuelling the likelihood of hot-dry events.
This nuance in the result was not initially planned and came as a consequence of a generous contribution by the reviewer. This made me appreciate a few more things about the process of setting up a research paper for publication. First, there is always going to be more things to investigate, it is important to know when to stop and send the preprint for revision. Second, science is full of enthusiastic and good people that are willing to invest time for rather selfless aims (i.e. betterment of research). Finally, there is often value in leaving yourself open for criticism. I hope the paper will reach out to other researchers interested in the impacts of compound events. Here we highlight that there is a need to study risk away from univariate extremes and more from a system perspective as damaging events are rarely occurring disconnected and in stand-alone mode. Future research should further expand the boundaries of that system to include important socio-economic drivers of yield variability.
Hamed et al. (2021) – Impacts of compound hot–dry extremes on US soybean yields – Earth System Dynamics