Introduction
The spatial modeling class taught us how to model a spatial process. This project models the movement of Hurricanes in the Atlantic Ocean. Scientists have developed models to track the path of hurricanes and continue to improve them. Over the years, the US model has been competing with the European model. The European model has proven to be more accurate at predicting hurricanes. Hurricanes are based on a variety of factors including winds, sea surface temperature, and rotation. The goal of this project was to create a simulation that tests the importance of sea surface temperature against pressure in the development and track of hurricanes. The group hypothesize was that sea surface temperature has a greater impact on the development and track of a hurricane because water retains energy far better than air, this allows for greater amounts of heat transfer to power the storm.
Methodology
The hurricane model was a group developed project. The work was divided into data collection, coding and documentation. Each of the group members had data layers to develop including sea surface temperature, pressure and land surface polygon. I developed most of the modeling script in NetLogo. The more complex NetLogo code snippets were built collectively. The main documentation was ongoing throughout development. We each had to write an individual portion as well. This can be read in the full report linked at the bottom.
The first step was to collect data layers including sea surface temperature, land cover, and pressure. All the data layers had unique challenges to overcome. The sea surface temperature had to be converted from weeks to daily layers. An interpolation script was developed to create the daily sea surface temperature layers. The pressure layer had to be converted from months to daily layers. A similar method was developed for the pressure layer to make the conversion to daily layers. The daily layers were used to realistically move the hurricane from day to day. With only one layer the hurricane would track the same path every run.
Sub-models were created to run the basic movement of the hurricane (Figure 1). A hurricane starting at the equator will generally follow the north east trade winds to the west. At higher latitudes the hurricane will follow the westerlies. The window patch was developed to model this occurrence as closely as possible. The window moves with the hurricane and switches above latitude 30N. That was our attempt to simulate global winds (Figure 2). The next step was to develop the hurricanes movement. This was done through the window patch, strengthen and dissipate functions. The functions “strengthen” and “dissipate” responded to sea surface temperature and land cover. The hurricane could move farther at higher strength. The strength would increase at high sea surface temperature. If the hurricane hit land then it would begin to dissipate. These were attempts to approximate the life span of the hurricane.
Results
The results of the model are shown below. Figure 3A and 3B compare the results of our model to the actual track of the hurricane Florence. It is not perfect but model follows the general path of actual hurricane pretty well.
Figure 4A and 4B compare the path of hurricane Michael. This was a great example of how hurricanes can react on land. The model failed to predict the continuation of the hurricane north. The model simply stops after reaching land. This is one of the limitations of the model.
Conclusion
Once the model was finalized, we compared the modeled hurricane tracks to actual hurricane tracks that season. Although there were a number of differences, the model did a fairly good job of modeling the actual hurricane tracks. Although we were not able to determine which factors were more influential, we concluded that hurricanes are extremely complex structures that are difficult to model accurately.
Skills
- Model Building
- Basic Programming and Scripting
- Project Management
- Communication
Melissa Gfeller 