Spatial probabilistic patterns of joint heat and water stress of chickpea in Australia
Quantitative characterisation of the crop environment for breeding and agronomic applications is commonly based on the cluster analysis of isolated climate factors, such as drought and temperature. The focus on individual climate factors fails to account for both the correlations between climate factors in space and time, and the contemporary models of dynamic biological systems where multiple interacting factors simultaneously shape the crop phenotype. With a focus on chickpea in Australia, our aims are to (1) assess the associations between climate factors and actual yield measured in 578 location -years; associations are investigated on a biologically relevant developmental scale; (2) establish the spatial, probabilistic patterns of multivariate environment types at regional and continental scales; and (3) evaluate the shifts in the frequency of environment types with climate change. We identified a syndrome of hot, dry and high vapour pressure deficit with three intensities of stress classified as low, medium and high. Measured yield, in the range from failed crop to 4 t ha-1, aligned more strongly with this multivariate syndrome than with isolated climate factors, with stronger associations at regional than continental scale. The frequency of stressful environment types increased with realised climate change and was most severe in the chickpea growing heartland of central Queensland; our models show that increased frequency of stress could be partially mitigated with earlier flowering varieties. Our findings will inform chickpea agronomy, phenotyping and breeding.