Mapping land cover in urban residential landscapes: Implications for water budget calculations.

摘要

Residential landscapes consume a considerable portion of states' water reserves. Landscape irrigation accounts for 40% - 70% of the household water use in the United States. For Albuquerque, the largest city in New Mexico, ground water is the main source of water and Per Capita Water Use (PCWU) was 167 gallons/day (632 liter/day) in 2007. This was more than most Southwestern major cities in the United States Albuquerque's water conservation programs are targeting a PCWU of 155 gallons/day (587 liter/day) by 2024. Thus, the major goal of my study was to obtain comprehensive knowledge about residential landscapes that are present in Albuquerque. This information will allow both homeowners and planners to effectively participate in water conservation efforts. To fulfill the goal of my research, I conducted two studies. In the first study, I classified residential landscapes of Albuquerque into categories and determined the spatial composition of residential landscape vegetation components. In the second experiment, I developed a landscape water budget web interface that allows homeowners and decision makers' calculate landscape water budgets. Development of the web interface demanded a wide range of information to assure accurate water budget estimation. This information included spatial and temporal information about reference evapotranspiration (ET0), existing residential vegetations and landscape plant coefficients. However, this information was virtually nonexistent.;In my first study, I resolved the complexity of residential landscapes in Albuquerque by categorizing (subdividing) its landscapes into distinct categories. I acquired comprehensive knowledge of greenspace (trees, shrubs and grasses) distribution among zip codes and among median incomes. I used Geographic Information System (GIS) vector files for Albuquerque parcels, city limits, zip codes and land use to create an integrated city parcels map. This map represented single and multiple family housing within the city boundaries. This information was incorporated with the ownership information data base obtained from the Assessor's Office of Albuquerque. Thirty residential parcels were selected randomly from each of the Albuquerque's 16 zip codes (strata). The object-based supervised classification module in ENVI EX 4.7.1 software was used to identify residential objects from very high spatial resolution (0.5ft2) true color aerial photographs taken in 2008. Overall classification accuracy was 89%. I identified eleven landscape categories in the heterogeneous landscapes of Albuquerque. Trees dominated landscapes of older and lower income zip codes, while grass primarily occupied landscapes of newer and higher income zip codes. Knowledge of residential vegetation components allowed landscape categories to be identified in each zip code or median income group.;In the second experiment, I utilized these residential vegetation components to develop mixed landscape coefficients for each zip code. After developing a Region Match Analysis, I used the non-turf landscape plant crop coefficients (Kc) developed for California Climate Region 2 and applied them to Albuquerque. Region Match Analysis included testing the level of matching weather conditions between different locations in California and Albuquerque. Monthly and yearly ET0 values were included for each zip code, because statistical analysis revealed significant difference in ET0 among zip codes. In addition, El Nino Southern Oscillation (ENSO) phases ET0 also were statistically different, and their ET0's were added to the web interface. The interface uses three water budget calculation approaches. First, an approach similar to the water budget method commonly was used, except that I included a ground-proofed vegetation coefficient for each zip code. Second, the Vegetation Fragmented Water Budget approach was used to calculate the residential landscape budget based on vegetation (trees, shrub and grass) areas and the associated general vegetation Kc. Third, Species Fragmented Water Budget used landscape areas of each species and specific Kc's for each species.;In the interface, available at this link: www.nmclimate.nmsu.edu/wb, users can access their residential landscapes' top view images to digitize and calculate vegetation areas and choose between different methods of calculation, ET0 bases, plant types, units, and irrigation efficiencies. Water budget calculations showed that the use of these approaches provide lower water budget estimates compared to common water budget approach. This web interface may serve as a critical tool for decision makers, planners or municipalities. My research has offered improved methods of water budget calculations, provided spatial and temporal ET0 values, calculated ground-proofed mixed landscape Kc and specific Kc's for lists of popular landscape plants. In addition, the newly developed interface offers the possibility of examining categories of landscapes existing in each zip code to target specific water conservation efforts.