Alec Hoffman
Lab 5
Agricultural Zone
1. To determine the percentage of guava in each farm, I used thematic raster summary (by polygon) in Hawth’s Tools. This gave me a dbf with the total number of 0 cells and 1 cells in each farm area. I joined this dbf to the farm shapefile and created two new columns, totals and percentages. Using field calculator I added the 0 and 1 fields to come up with a total and for percentages I divided the number of 1’s by the total. I looked at various different graphs, but there was no clear trend, until I created another new field, total operations. This was the total of the cattle, crops, and pasture fields. The graph in the folder shows that there is a trend towards a smaller percent of guava as the total operations increased.
2. The map clearly shows that around most of the buildings there is a lower percentage of guava than in other fields. The areas with the greatest concentration of guava are not close to any buildings and the guava percentage is reduced as one moves closer to the buildings. The connection between guava and roads is not so clear, however. In areas were the roads are close to buildings, the guava percentage is much less than other areas, but along the roads in areas far from buildings the results are more mixed, some areas have little guava while other areas have a large percentage of guava.
3. I added a new field, Status, and started editing. I determined that any farm that was below average on guava and above average on operations was active, any farm that was above average on guava and below average on production was abandoned and any farm that either was about average in both areas or had above average production and guava, or below average production and guava was partially abandoned. I used the statistics tool to determine the averages for the categories. I created a map that shows the different status of the farms.
4. Clustering seems to occur around farms where there are other activities, such as fishing and tourism. Since these activities do not depend on clear fields there is probably no need for the people to expend the effort to remove the guava growing on the farms. Cattle and pasture also seem to be associated with areas of little guava, possibly the cattle are grazing on the guava trees when they are young, preventing them from growing, or the farmers are forced to remove the guava to give their cattle enough grazing land.
Park Zone
1. The calculations show that the number of guava pixels decreases dramatically as you get farther away from the park boundary. The initial ring of 600 meters has 582,064 pixels and that falls off to 34,252 guava pixels at the outer most ring.
2. Comparing the out of park pixels with the farm fields, the only thing that seemed to stand out was the size of the fields. Larger fields tended to be associated with more guava pixels in the park. I do not think that it is very significant, though, because larger farm fields with have larger areas attributed to them. Otherwise, I did not see any correlation between specific information about fields and guava pixels in the park.
3. Comparing the count of pixels with the other fields in the farm database did not show any clear patterns. There were active and abandoned farms as well as partially abandoned. There did not seem to be any correlation between the area of the farm and pixels or the amount or type of actives on the farms.
Summary
Based on the data provided, there seemed to be two clear factors that determined the amount of guava in the area. The first factor is the status of the farms, active farms had much less guava in terms of the overall land area compared to abandoned or partially-abandoned areas. The second factor is how close parkland is to farmland. The park areas within a few hundred meters of the farm boundary had significantly more guava, up to 10 times more, than park areas farther away from the farm boundary. This factor did not seem to be affected by the type, size, or status of farm that bordered the parkland, it was uniformly overrun with guava. This could be due to the fact that parkland on the farm boundary is equivalent to abandoned farmland, it is close to affected areas and there are not any people willing or able to remove the guava.
Although it seems counterintuitive, it might be useful to bring a few people to work the abandoned farms again as a way to reduce the amount of guava in the farm area and prevent further growth. It is probably not wise to expand the farm area deeper into the park, because that would accelerate the spread of the guava into the park area. If feasible, hiring people to clear out the guava could be a possible solution.
To have a better idea of the problem, time elapsed data would be helpful. The speed that the guava is spreading out into the park will determine how quickly action needs to be taken to stop the growth. Other data that would be helpful is more accurate farm information, because it seems that farms have spread beyond the boundary. IF that is the case, it would make more sense to relocate those farmers back into the boundary area in abandoned farms. Also, there was not any data on animal movements and tourism patterns, because both of these could be exacerbating the guava spread.