Australia’s largest pesticide monitoring program (the Great Barrier Reef Catchment Loads Monitoring Program) is conducted as part of the Reef 2050 Water Quality Improvement Plan (WQIP). While this program targets the largest and potentially high risk rivers the resulting monitoring data can only inform us of what is happening at the sampling points. The WQIP has a target for reducing pesticides. This was initially a loads-based target but was recently changed to a risk-based target (i.e. to protect at least 99% of aquatic species at the end of catchments from harmful effects of pesticides). Thus, it was necessary to develop a method that could estimate the toxicity of pesticide mixtures and develop models that could predict pesticide mixture toxicity where no monitoring data exist and at multiple spatial scales (catchments, basins, regions and the Great Barrier Reef Catchment Area). In this project we developed a multi-substance potentially affected fraction (msPAF) method and combined that with the response addition model of joint toxicity and multiple-imputation to estimate the average daily toxicity of water samples, containing up to 22 frequently detected pesticides, over the wet season. This estimates the toxicity of pesticide mixtures at the sampling site, but such information is needed for other waterways and different spatial scales. So relationships were developed using land-use, hydrological and site-specific parameters that could accurately predict pesticide mixture toxicity. These relationships were initially used to estimate the pesticide mixture toxicity of 35 basins, 6 Natural Resource Management Regions and for the entire GBRCA. Subsequently, the relationships have been used to predict the toxicity of 170 catchments of varying sizes and we are currently using them to predict the toxicity of pesticide mixtures in every 1 to 1.5 kilometre stretch of Queensland waterways.