Mapping mean monthly temperatures over a coastal hilly area incorporating terrain aspect effects

Efforts in the past two decades on air temperature mapping based on sparse monitoring networks reveal that algorithms based on multiple linear regressions with geographical and topographical parameters perform promisingly. In this study, a multiple-regression model, previously for precipitation characterization using autosearched orographic and atmospheric effects (PCASOA), is applied to analyze spatial distribution of mean monthly daily maximum and minimum temperatures (at 33 stations) in Adelaide and theMount Lofty Ranges (9000 km²), a coastal hilly area in South Australia. Terrain aspect (or slope orientation) is transformed and explicitly incorporated in the model, together with some other topographic variables. Overall, PCASOA captures 91% and 70% observed spatial variability for mean monthly maximum (T_max) and minimum (T_min) temperature, respectively. The regression also infers some physical processes influencing the air temperature distribution. The results indicate horizontal gradients of T_max in the east-west and north-south directions,which can be related to the effects of dominant wind directions in the study area. The effect of terrain ruggedness on T_max is likely related to the blockage of sea breeze in the complex terrain. Cold air drainage potential only influences T_min during winter months in the study area. Terrain slope and aspect significantly contribute to interpreting T_min spatial distribution and can be related to their sheltering effect fromthe dominant cool inland winds. They also contribute to interpreting T_max spatial distribution, while the physical mechanism is not clear.