This is a data intensive assignment that will require a time investment on your part.

The main database that you will work with is located HERE

You will be preparing an electronic report that includes all relevant graphs, tables and statistics. Everything should be in one PDF document. That document, when completed, should be uploaded via The Homework Submission Form

Part A:

Your mission is to use the various maps that you can configure and make on different spatial scales (individual states, regions, etc) to identify large scale temperature and precipitation anamolies. These analomies are defined as having either very high numbers (usually encoded in red) and/or very low numbers.

You are to identify as many of these events as you can find. By way of example:

• Set the year to 2012
• Set the month to April
• Set time span = 3 months
• Set product to State Temperature Rank

Those settings should produce this map:



Note that by holding down the control key you can select many products at once - this will help you identify appropriate event maps that you judge to represent large scale departures from normal. You are to do this for both TEMPERATURE and PRECIPITATION and include all interesting maps in your report.

Identify any systematic regional trends that you discover and include that in your report.

Part B:

There are many ways to do this part. You are to develop some kind of statisical approach to represent the frequency of occurence of rare/record events. For instance, let's take the map generated above:

There are 118 years of data. If everything was normal then the mean value on the map would be 59 with some spread around that due to random climate noise. So there is some probability associated with, given that mean expectation, there are N events of say amplitude 118 or N events above 110 or something like that. The simple observation that every state is above 59 in this map is important. There are many ways to do this statistically. You could try to apply Poisson statistics, or you can just use a random number generator model to assess probabilities. Of you can make toy models where increases are expected with time. The entire point here is:

a) You have identified deviations from normal visually using the various maps that you have generated

b) now you want to develop some framework to establish the statistical signifcance of this much deviant data occurring since 2002 (when the database starts).

Note: there is no "right answer" here, there is only quality of research your doing to try and demonstrate that regional climate change is occurring, perhaps at an alarming rate, in various places in the US.

Note: You are to work only with the RANKS maps and not the other products that are available.

It might be possible to use poisson statistics on the rank number. here is some information on these kinds of statistics