Methods and Study Area

The study area is eastern Bernalillo County, New Mexico, chiefly the Sandia Mountains and the surrounding area, including the parts of the City of Albuqueruqe, the Pueblo of Sandia, the Village of Tijeras, and the foothills of the Manzano Mountains. Four weather stations with sufficient data to compute ET using the Kimberly-Penman method were identified. Two of the weather stations belong to the Remote Automatic Weather Stations (RAWS) network, which analyzes meteorological data to compute fire danger ratings throughout the country (https://raws.dri.edu/). These stations include Sandia Lakes, located in the Rio Grande Bosque riparian ecosystem in Sandia Pueblo, and Oak Flats, located on US Forest Service land southeast of Tijeras, NM (Fig. 1). The other two stations are part of a meteorological monitoring network run by the US Geological Survey, and they include Sandia Mountains Upper Precip Site (Site 1C), located near the Sandia Crest, and Tijeras ET, located in Tijeras, NM (www.usgs.gov). The elevations of the stations range from 5000 ft above sea level at Sandia Lakes to 10,091 ft above sea level at Sandia Mountains Upper Precip Site (Table 1). The stations are located in various ecosystems: a cottonwood-dominated riparian woodland (Sandia Lakes), a Pinyon-Juniper woodland (Tijeras ET), and Ponderosa Pine-scrub Oak woodland (Oak Flats) and a mixed aspen-coniferous forest (Sandia Mountains Upper Precip Site). The ET values for the weather stations were calculated using the Kimberly-Penman equation (Appendix A).

Table 1. Weather Stations

Site_name	Latitude	Longitude	Elevation_ft	Agency
Sandia Lakes	35.23	-106.591	5000	BIA
Oak Flats	35.00417	-106.322	7550	USFS
Sandia Mtns Upper Precip Site (Site 1C)	35.20803	-106.435	10091	USGS
Tijeras ET	35.07624	-106.384	6366	USGS

Figure 1. Study Area with Weather Station Locations

For comparison, MODIS 500-m 8-day composite evapotranspiration data was obtained from the NASA EarthExplorer website (search.earthdata.nasa.gov). The evapotranspiration from each 8-day period, beginning on the first day of each year, was expressed in units of kg/m²/8-days in a 500x500m pixel raster grid. Using ESRI ArcMap, erroneous data was nullified and then filled in (Figure 2). The evapotranspiration data calculated from the weather stations was then summarized into 8-day totals to match the MODIS 8-day composite grids, and a separate raster surface was created for each 8-day period (Figure 3). Then, the raster calculator tool in ArcMap was used to calculate the percent difference that the MODIS data deviates from the weather station data for each cell in each raster image. The GIS workflow is described in detail in Appendix B.

All layers in the map were projected into North American Datum of 1983 (NAD 1983). NAD 1983 is the datum to which the coordinates for the weather stations are referenced to. It also provides an accurate visual representation of the study location and it is the datum that is most familiar to authors of similar studies.

Figure 2. Landsat 8-day composite ET data, with erroneous data nullified and filled

Figure 3. Raster surface of ET derived from weather stations, 2/10/2014 – 2/17/2014