Investigating Vegetation in the Phoenix Basin

Data used for modeling:

The pilot study used several different sets of data, chosen for possible relevance to varying levels of biomass, or vegetation health. These included information on soil salinity, alkalinity and clay content, as well as vegetation communities. Using a principle components analysis, I was able to determine that these data added little information to the total set of data, and regression analysis showed that they had little effect on biomass. Other layers of data that were discounted for the same reasons are slope, and to some extent, aspect. (Aspect is taken into account under the solar intensity model I used, discussed below.) One significant data layer, derived in a Geographic Information System (GIS) from elevation values for Arizona, is flow accumulation. This is a measure of drainage, the total area draining into each location on the landscape. This is shown in the figure below.

Another significant data layer, again produced in a GIS from elevation values, is intensity of afternoon/evening sun. Based on my own experience with growing plants in the Valley of the Sun, as well as clear patterns in the biomass data, I realized that plants with afternoon and evening shade are much healthier during the summer. Because the biomass data are taken from July satellite images, the plants in the desert showed a similar pattern. Vegetation on east-facing slopes is much healthier. The image below shows a simulated summer  solar intensity value, produced by averaging a series of "hillshades" calculated from the elevation data from different sun directions and elevations for a late July day. The dark blue colors represent lots of sun, while the yellow-green colors represent shaded areas, where we also see healthy vegetation.

Shown in the figure below is the actual elevation layer itself, that was used to derive the two data layers illustrated above. Elevation itself is a smaller component of vegetation health, in the sense that we cannot predict biomass from a simple elevation value. However, there is still a small component of vegetation health related to elevation, perhaps because of the different vegetation types that are distributed with elevation. The low-lying areas are dominated by Creosote-Bursage communities, while higher elevation areas are dominated by Palo Verde- Cactus communities. The large mountain mass near the center of the study area is the Estrella Mountains. The smaller mountains in the extreme northeast are in South Mountain Park, while the scattered hills in the west are the Maricopa Mountains.

The data illustrated below were perhaps the most difficult to generate. This shows the averaged yearly precipitation values for 1986, 1992, and 1993, the years that inexpensive satellite data are available to me. The precipitation data was generated from rain gages across Arizona, collected into a database and maintained by the National Oceanographic and Aeronautic Administration. I used a geostatistical process called "kriging" to generate smooth precipitation maps from widely scattered rain collection points. This average precipitation map incorporates different rainfall patterns from a dry year (1986), an average year (1992), and a very wet year (1993).