I am using Cell Statistics (Spatial analyst/Local) to calculate mean values for a set of 32 bit floating point rasters; all the same size and source (a NetCDF), identical cell size, and numbers of rows and columns. The resulting overall mean raster has high values that are greater than ANY of the input rasters for which I want the mean value. All of the high values are near or somewhat near NoData values (but some are in the center of Florida, not near the raster edges). I've tried checking and unchecking the ignore no data option. Any help with this? I've seen some older posts about a bug that 10.4 has addressed, but that is not an option for me. Thanks.