# Identifying geographies that match certain climatic profiles through GIS

The long and short of this research is trying to find roads that lie within geographic areas that fit a given climatic profile. In this case, I need to find areas within the United States with relatively cold falls, winters, and springs that get snow. These places may not have a lot of snowfall on an annual basis, but the snow persists for longer periods due to the colder weather patterns.

Are there any websites that already collect similar data in a format ready for GIS analysis? If not, what should I look for when gathering data? There may not be a singly perfect answer for this question, but I mostly am seeking direction to refine my search appropriately.

• What kind of scale are you analysing? National? City wide? World-wide? May 12, 2011 at 20:08
• @geographika This is going to be done nationwide for the US May 12, 2011 at 20:13

A quick Google has turned up MODIS Snow and Ice Project, which appears to give resolutions of 500m.

You could also look at Landsat data. Due to the high albedo of snow, it should be a fairly easy process to threshold highly reflective values, then average them out by collecting a time-series of images and applying map algebra.

Alternatively, you could acquire weather-station data from somewhere like NOAA and use some form of interploation like Kriging to form a raster.

From there you could keep it as a raster layer and use something like PostGIS Raster to do intersection tests, or convert the raster into polygons with say, GRASS, and do vector-based tests.

For the United States, National Snow Analyses. For global try Rutger's Global Snow Lab

These were the top results for a Google search for "snow cover" and "global snow cover", respectively. You'll have to come up with a quantitative definition of what you're looking for (e.g. has snow cover of depth x over period y) and apply it to the data. Map algebra should come in handy.

• +1 sounds like classic map algebra type problem, then overlay the roads May 12, 2011 at 20:08

After some more digging, I found that NOAA and NCDC keep a respectable number of freely available datasets detailing just the kind of information I'm looking for. I was able to find shapefiles for:
1. Average Mean Temperature
2. Mean Number of Days with Temperature 32 degrees Farenheit or below
3. Mean Number of Days with Snow Depth >= 1, 5, or 10 inches.

I admit that stopping my analysis at this point leaves it rather shallow. I would love to do a deeper search that involves some of the methods mentioned by MerseyViking and Sean, but a) my experience with map algebra and climatography are very limited, and b) time is even more limited for results! Luckily, this will provide the information needed, and I thank everyone for their input. I will certainly keep it all in mind for the future.

I've worked with a few local government GIS departments in Europe, and many of them have data on when roads are de-iced, and the routes the gritting lorries take (as attributes of a GIS road network). All local government agencies should have some sort of figures on the budgets for maintaining roads in winter (salt, grit, gritters etc.)

After doing your analysis using the raster methods suggested by others, you could try and calibrate / validate your results by getting this data for a specific area and seeing if the results match your expectations.

If not, then you'll probably have to look again at your analysis!