This question is an extension (next step) to Comparison with different resolution rasters.
Currently I have successfully compared the similarity between the two raster datasets (currently both rasters have the same resolution), one from the PCA model, and the other from the VIC model on a daily as well monthly bases. The two tests that I used are the Kolmogorov-Smirnov (KS) and the Kruskal Wallis tests (KW). The results are good, but they don't tell me the spatial similarity between the two datasets i.e. how similar the two datasets are at each pixel? I have attached a sample dataset here. Grid Code is the pixel ID, and for the sample data, I am comparing "VIC_SC_04" with "PCA_SC_04". The data ranges from 0 to 1. 0 being 0% snow cover and 1 being a 100% snow cover. Note that during the end of the melt season, there are significant ties (0) in the data, as most of the snow has melted off.
Spatial data which is information about a particular location in space, and similarity is the analysis of divergence from equality (Bruns & Egenhofer, 1996). Moreover, spatial similarity according to Holt and Benwell is explained as areas at a specific resolution and attributes that are evaluated to be comparable or not. Thus, spatial similarity is dependent on spatial equivalence and classification based on the reason/purpose of comparison, and can also depend on a temporal facet of the data (Holt, 1999). What tests can I perform in R to do "spatial similarity analysis" along with plots between the two datasets on daily/monthly basis?
