ENNEC 472: Quantitative Analysis in Earth Sciences

Spring 2009

PROBSET #2 (DUE: Th Feb 26 at start of class)

Hypothesis Testing

1. Analysis of El Nino influence on Atlantic Landfalling Hurricane Totals

Follow up on problem #1 of problem set #1, by determining whether there is a statistically significant influence of the El Nino/Southern Oscillation (ENSO) phenomenon on landfalling Atlantic Hurricanes. Using the same data, determine if there is a significant difference between El Nino years and normal years? Between El Nino years and La Nina years? Between normal years and La Nina years? Be clear about what significance threshold level you’ve chosen for the p value? Did you invoke a one-sided or two-sided hypothesis test? Explain.

2. Analysis of Changes in Landfalling Atlantic Hurricane Activity

It is often stated that there was a ‘shift’ to a more active regime of Atlantic Hurricane activity beginning in 1995. Examine this claim using the long-term Hurricane data from problem #1 above, employing an appropriate hypothesis test. Be clear about what significance threshold level you’ve chosen (i.e. the p value) and whether you have invoked a one-sided or two-sided hypothesis test, and why.

3. Analysis of Data Distributions

Assess the degree to which the following data series analyzed in problem set #1 conform to a “Gaussian” distribution:

a. Indian Monsoon precipitation

b. State College December Temperatures

You may make use of the Matlab subroutine “gaussparent.m” used in class, but you should understand what the routine is doing [you will also need to download the subroutine “quantiles.m” which is called by “gaussparent”].

4. Analysis of Changes in State College December Temperatures

Suppose that climate model simulations indicates that December temperatures in Pennsylvania should have become both (a) warmer and (b) more variable during the latter half of the available record as compared with the earlier half. Use appropriate hypothesis tests to address the extent to which the observations support the model predictions. Be clear about what significance threshold level you’ve chosen (i.e. the p value) and whether you have invoked a one-sided or two-sided hypothesis test.