The environment is the largest component of the phenotypic variance of crop yield, hence the importance of its quantitative characterisation. Many studies focussed on the patterns of water deficit for specific crops and regions, but concurrent water and thermal characterisations have not been reported. To quantify the types, spatial patterns, frequency and distribution of both water stress and thermal regimes for chickpea in Australia, we combined trial and modelled data. Data from National Variety Trials including sowing time, yield and weather from 295 production environments were entered into simulations. Associations between actual yield, in a range from 0.2 to 5.2 t/ha, actual temperature and modelled crop water stress were explored. Yield correlated positively with minimum temperature in the 800 degree-days window bracketing flowering and the correlation shifted to negative after flowering. A negative correlation between maximum temperature over 30°C and yield was found from flowering through to 1000 degree-days after flowering. Yield was negatively correlated with simulated water stress from flowering until 800 degree-days after flowering.

Cluster analysis from 3905 environments (71 locations × 55 years between 1958 and 2013) identified three dominant patterns for both maximum and minimum temperature accounting for 77% and 61% of the overall variation, and four dominant patterns for water stress accounting for 87% of total variation. The most frequent environments for minimum and maximum temperature were associated with low actual yield (1.5–1.8 t/ha) whereas the most frequent water-stress environment was associated with the second lowest actual yield (1.75 t/ha). For all temperature and water-stress types, we found significant spatial variation that is relevant to the allocation of effort in breeding programs.