Conclusion

Precipitation 

The krigs created from the data in the weather stations seemed to have much more precipitation and much greater spatial variability.  For instance, the maximum average precipitation estimated by the PRISM data was around 50 mm, while the maximum precipitation shown by the weather station data was around 97 mm.  

There are a couple reasons that this may be, both of which are rooted in the fact that the precipitation data is from a different time period than the PRISM estimated data.  The weather station data was from 1997-2009, while the PRISM data was from 1971 - 2000.  The greater spatial variability in the krigs can be explained by the fact that the PRISM data was over a greater period of time, thus allowing the solution to be much more smooth.  The krigs showed greater precipitation, and one reason for this possibility that can be discarded with almost complete certainty is that the rain gauges overestimated the amount of precipitation.  Rain gauges, if they are inaccurate, are generally underestimating the amount of rain.  Reasons for the discrepancy may be that there was more rain in the later years.  Climate change has caused a shift in global temperature, giving air more capacity to store moisture, which may explain increased precipitation.

Maximum Temperature

In most months, the result was the same, namely that the PRISM data had much greater spatial variability and in general, higher temperatures.  The reason that the weather stations estimated lower temperatures than the PRISM data could be that the data gathered by the weather stations could have been slightly underestimating the temperature.  The reason for the spatial variability in the PRISM data could be that PRISM group incorporates point weather data, elevation, and other climatic phenomena that may influence the temperature.  The krigs, although there is more point data, do not take anything else into account; the rasters created are simply the result of a mathematical calculation whose parameters are simply the values of the point data. 

Based on these considerations, a logical conclusion to come to is that the PRISM data, while correct in its estimate of spatial variability, may be overestimating the actual maximum temperatures.  The incorporation of considerations other than point data make the spatial variability seem logical, and the greater amount of weather data add to the credibility of the lower temperatures.