The analysis models I developed for this project are preliminary. My primary goal was to generate a logical framework of spatial analysis steps and to explore some of the ways in which the data available in SWERA's Nicaragua Geospatial toolkit can be used to evaluate locations as potential sites for wind energy generation. The need to complete this project in a single trimester meant that there are many other potential data sources and analysis steps which could be used to refine this process that I did not have time to explore.

The most important next step is to conduct more detailed research to confirm or refine the basic assumptions and criteria of the models I have created. An example is the criteria used for maximum slope. I set the maximum slope as 30 degrees because this seemed steep enough to pose a major obstacle to installing wind turbine towers. Future research into the engineering requirements of different turbines may lead to a revision of this value. It is also possible that with more detailed information I might learn that the maximum slope should be different for different sizes or models of turbine, or perhaps with a better understanding of the actual process of installing towers on different slopes I would determine that a ranking model for slope would be more appropriate than a binary model.

The criteria for distance from a community also can be refined. In the SWERA data, communities are represented by a point feature class. In the future, information about the population and area of each community can be used to refine the criteria for distance from community. In my model for off-grid turbines I used a buffer distance of 500 meters, longer than would likely be practical to run electrical wire to a small turbine, because the physical size of the community is not known. The maximum acceptable distance from the center of a community for an off-grid turbine might vary by the area of the community, its population, the quality of the wind resource available or perhaps a combination of more than one of these factors. Similarly, the area around each community that should be excluded from commercial-scale wind production would vary with population and area as well.

The criteria for aspect can be refined. First, instead of using a binary model, a ranking model could be used to assign ranks to slopes of different aspects. Another limitation of the current models is that the aspect criteria are based on wind rose data sampled in Bluefields. There is room for debate as to how large an area is reasonably well represented by this data but it is obvious that other parts of the country will require different aspect criteria to adjust for prevailing wind conditions. One solution would be to find wind rose information for a sample of locations across the country and base the aspect criteria for a location on the closest available wind rose data site.

I chose to omit the distance from transmission lines as a criteria for off-grid wind power because I do not have enough information about the availability of transformers or which communities currently receive power from the grid. Proximity to the transmission line might not be enough to ensure that a community has grid power available. Additional information about the electric infrastructure and electrification would allow this evaluation to be refined to exclude communities that already have adequate electricity. The population and economic data about communities could also be considered to forecast electrical demand and ability-to-pay to address the financial suitability of different sites.

As a ranking model is used to process more factors, the issue of how to weight these various ranks will have to be addressed. This is where the work of validation, comparing actual measured results with predicted results, will have to be addressed to assure that the models are as useful as they can be for their intended purpose.