Degree Type


Date of Award


Degree Name

Master of Science


Geological and Atmospheric Sciences

First Advisor

Eugene S. Takle


Our ability to numerically simulate near surface winds is a challenging, yet necessary component within meteorological and climate models. There are many societal implications of winds both high and low such as wind energy, air pollution dispersion, agricultural productivity, as well as the sometimes catastrophic damage to life and property. Due to these impacts, we seek answers to the question of how well our weather forecast and climate models with low resolution simulate near surface winds. We also seek to know how these models simulate surface winds into the future with respect to climate change.

We evaluated the characteristics of three regional climate models forced by NCEP reanalysis II data across five locations in the Midwest, United States: Mason City, IA, Lafayette, IN, Hastings, NE, Liberal, KS, and Jamestown, ND; and two southern metropolitan cities: Houston, TX, and Atlanta, GA. The first part of the analysis was carried out to provide a baseline to test these climate models' ability to accurately simulate surface wind conditions using observational data as a reference point. There is a general negative bias in both the climate models with NCEP reanalysis data as well as the contemporary climate models at all of the locations except Houston, TX, and Atlanta, GA, which had a positive bias. It also appears that the CGCM3 GCM introduces significant error into the contemporary scenarios at four of the seven locations. These are factors to take into account while formulating conclusions on the accuracy of the future scenario trends as well as the overall comparison between the future and contemporary climates.

Contemporary (1968-2000) and future (2038-2070) scenarios simulated by these regional climate models were also evaluated. Both low and high ends of the "extreme" wind spectrum were analyzed, in which our low-end "extreme" threshold is defined to be at or below the 10th percentile, and the high-end extreme to be at or above the 90th percentile. Seasonal distributions were also evaluated amongst each of the climate models. Overall, the contemporary and future scenarios appear to simulate the general timing of seasonal minimum winds (June, July, and August), whereas, they do not simulate seasonal maximum winds with accuracy (March, April, and May).

When comparing the difference between future and contemporary scenarios, it is evident that near-calm winds show to be increasing in frequency across all of the stations analyzed and high-end winds are showing inconclusive trends throughout the climate models studied. The MM5I displayed an anomalously high frequency of low-end winds at six out of the seven locations compared to the other regional climate models. This unusual feature needs to be further investigated because of air pollution dispersion and agricultural implications. There is large variance among the climate models, so it is recommended to exercise caution when using a single model for applications or references.

Copyright Owner

Rachel Hatteberg



File Format


File Size

135 pages

Included in

Meteorology Commons