Effects of spatial resolution on predicting the distribution of aquatic invasive species in nearshore marine environments

The most appropriate range of spatial resolutions of environmental data with which to accurately delimit potential distributions of aquatic invasive species (AIS), in shallow nearshore marine environments, using species distribution models (SDM), is currently unknown. This study used SDM to determine the optimal range of spatial resolutions of temperature and salinity data with which to predict the potential distribution of vase tunicate Ciona intestinalis on the Canadian east coast and European green crab Carcinus maenas on the Canadian west coast. Both of these problematic AIS have spread rapidly in temperate nearshore coastal waters. As the invasion success of these species in temperate seasonal environments is constrained by temperature and salinity, we used SDM, specifically MaxEnt, to correlate these environmental variables at a range of spatial resolutions (100 km to 100s of metres, the latter encompassing 100 or 500 m(2) on east and west coasts, respectively) with both species' occurrence data. Increasing spatial resolution from 100 km to 100s of metres of temperature and salinity data generally resulted in more accurate estimates of each species' distribution, including a more realistic depiction of how salinity and temperature shape their distributions, with several caveats. First, increasing resolution of temperature and salinity data did not translate into proportional increases in model performance. Secondly, the highest resolution (100s of metres) did not result in the most accurate predictions of east coast C. intestinalis distribution. Finally, lower spatial resolutions (i.e. 100 km to 8 km resolution) performed worse in MaxEnt for west coast C. maenas than for east coast C. intestinalis. Overall, finer-resolution patchiness in each species' distribution was accurately resolved at or below spatial resolutions of 9 km for east coast C. intestinalis or 4 km for west coast C. maenas.