Australia is a dry continent and therefore irrigation is an essential activity to the agricultural sector of the Australian economy. Irrigation requires water diversions that consequently alter stream flows, which are highly contentious. Water supply is also affected by weather and climate variability, with events such as the Millennium Drought (2001-2009) having significant consequences, particularly on the ability to maintain irrigation activity in the Murray Darling Basin (MDB), one of the main agricultural regions of Australia.

Irrigation has been shown, using both observations and climate models, to affect the climate through localised cooling, particularly of hot temperature extremes. However, most existing analyses has utilised data or model output at spatial resolutions in excess of 50km. In the global context, this resolution is too coarse to resolve the irrigation effect for regions such as the MDB where the intensity of irrigation is not as great compared to India. Using high-resolution datasets, we apply a window-searching algorithm to reconstruct the effect of irrigation on the climate of the MDB over the period 2003-2017.

We estimate the uncertainty by applying the algorithm to several datasets including:

• ANUClimate daily mean temperature product at 0.01˚
• AGCD daily mean temperature product at 0.05˚
• MODIS monthly mean land surface temperature at 0.05˚
• BARRA daily mean temperature at 0.11˚

Our analyses reveal that irrigation activity in the MDB has contributed to localised cooling of the order of 1˚C for mean mid-day temperature. However, there is considerable variability in how much cooling occurs linked to water availability. During the latter half of the Millennium Drought, irrigation activity decreased due to severe water shortages, which had consequences for the cooling potential of irrigation. In the context of the current drought, irrigation activity is again likely to reflect changes in water availability with reduced irrigation likely to contribute to enhanced temperatures.