Josh S. Brown
Introduction
ArcET allows for a GIS based estimation of
evapotranspiration (ET) to provide a way to process data inputs for spatially
distributed information. An extension to
ArcGIS, ArcET, was developed to calculate large non-point specific evapotranspiration
for better estimates of water demands at regional levels. An accurate method for determining the
agricultural water demand is critical due to the increasing water demands in
portions of the
Evapotranspiration
Evapotranspiration (ET) is the combination of two separate processes, evaporation and plant transpiration. Evaporation accounts for the movement of water to the air from sources such as the soil, pavements, wet vegetation, rivers, lakes, streams, and other water bodies. Evaporation of water requires relatively large amounts of energy, either in the form of sensible heat or radiant energy to occur. Plant transpiration accounts for the movement of water within a plant and the loss of water as vapor through its leaves. Transpiration and evaporation depend on the energy supply (solar radiation and ambient air temperature), vapor pressure gradient and wind. Therefore, radiation, air temperature, air humidity and wind terms should be considered when assessing transpiration.
The evapotranspiration rate is usually expressed in millimeters per unit time. The rate shows the amount of water lost from a crop surfaces in units of water depth. The time unit can be an hour, day, month, a growing period, or year. Evapotransiration in terms of water depth can also be expressed in terms of energy received per unit area. The energy refers to the energy or heat required to vaporize water. This energy is called latent heat of vaporization and is a function of the water temperature and is expressed in units of MJ/m2 per day. The units used to represent ET that will be used for this term report will all be a monthly average reported in mm/day.
As briefly mentioned above, a combination of the weather parameters and crop characteristics are the driving forces behind evaporation and transpiration. The weather parameters affecting evapotranspiration are radiation, air temperature, humidity, and wind speed. The evaporation power of the atmosphere is expressed by the reference evapotranspiration (ETo). In most cases, grass (yes, the same plant that commonly provides most people a lawn) is used as the reference crop and is the basis of comparing crops with one another as a reference point. Crop evapotranspiration (ETc) refers to the evaporating demand from crops that are grown in large fields under optimum water amounts in the soil to get full production under the given climate conditions (Allen et al. 1998). Figure 1 shows the relationship between the reference crop evapotranspiration and crop evapotranspiration.
Figure 1: Shows the difference of ETo and ETc and how they are calculated.
Evapotranspiration Estimation
Methods
Evapotranspiration is not easy to measure. Evapotranspiration is calculated by measuring the different parameters of the soil water balance. This method consists of assessing the amount of water leaving and entering into the crop root zone over a specific time period where irrigation and rainfall add water to the root zone. Specialized devices called lysimeters (an instrument for determining the amount of water-soluble matter in soil) and accurate measurements of several physical parameters of the soil and water balance are required to determine evapotranspiration accurately (Allen et al. 1998).
Another method of estimating evapotranspiration is the mass transfer approach. This approach considers the vertical movement of small particles of air above a large surface. Evaporation from an open water surface (pan evaporation) provides an index of the overall effect of radiation, air temperature, air humidity and wind on evapotranspiration. The methods outlined above are often very expensive, extremely difficult to get measurements that are accurate and should only be preformed by well-trained personnel and often provide climate specific results that are not easily transferable from one region to another (Allen et al. 1998).
Several empirical to semi-empirical equations have been developed to determine evapotranspiration. These empirical equations require input that range from two or three parameter as simple as elevation and temperate of a region to very complicated equations that require many steps to find the evapotranspiration parameters. Most methods are derived for specific climates and need to have some form of calibration to find accurate estimates for evapotranspiration values. Below, in the ArcET extension, lists the most popular empirical evapotranspiration equations and located in the results section and their respective equations. Because many crops or region’s climate can be accurately estimated based upon elevation in combination with local temperature data, ArcET uses interpolation to better estimate regional climate parameters over potentially large areas.
ArcET Extension
ArcET is readily available for download at a website developed by Shujun Li at the following link (http://hydrology.neng.usu.edu/arcet/). Shujun Li was also the developer of ArcET. The user’s manual can also be downloaded with ArcET. The ArcET extension calculates evapotranspiration by using sophisticated ways to interpolate meteorological parameters collected at site-specific points in a region. The data that is interpolated are temperature, humidity, solar radiation, and wind. ArcET uses the five main empirical equations that are listed below.
• FAO 56 Penman-Monteith
• Standardized ASCE Penman-Monteith
• Hargreaves 1985
• SCS modified Blaney-Criddle
• Priestley-Taylor
ArcET Required Inputs
The required inputs for determining evapotranspiration, using the above listed equations in ArcET range from tables to grids. ArcET pulls all the spatially distributed data and relates it according to the relationship that the five methods outline. The following data inputs are needed to calculate crop evapotranspiration:
• Weather Stations – Location of weather stations.
• Weather Data Table – Weather information at weather station (i.e. precipitation, wind speed, humidity, etc.)
• Digital Elevation Model (DEM) – To get elevation data for the area of interest to interpolate weather information.
• Land Use/Cover Shapefile – An ESRI polygon shape file for crop distribution information
• Crop Coefficients Table – A DBF table with crop coefficients for each crop type in region
ArcET was developed to interpolate grids of spatially distributed meteorological parameters from point measurements. From these interpolations, reference evapotranspiration is then calculated at each grid cell. Next, with the crop coefficient grids and land cover/use information inputs, crop coefficient grid maps are generated. Then, from the crop coefficient maps and the reference evapotranspiration grid, the crop evapotranspiration map is developed. This relationship that is described above is laid out in Figure 2 below.
Figure 2: Shows how ArcET uses input data to generate crop
evapotranspiration maps.
The weather stations and weather data table information was collected from the National Climate Data Center (NCDC) website under the “Map / Data Access Services.” The data collected from the weather station locations gave the longitudinal and latitudinal components to show the spatial distribution of the weather stations, this data is also used in some of the equations to estimate of solar radiation and the length of sunrise for the given time of year (when solar radiation data is not available). The data collected for the weather data table included data that gave the mean high and low temperature for each of the 12 months, along with the average monthly precipitation for 2005. Aside from the obvious calculated parameters, these data sets where then used to determine estimates for vapor pressures. ArcET requires a minimum input of only the temperatures for each weather station. All other weather parameters can be interpolated (like precipitation) and parameters such as wind speed can have default values used as input.
The fundamental data input using ArcET is the digital elevation map (DEM) for the area of interest. DEMs can be downloaded from the United States Geological Survey (USGS) website, although a DEM of Utah that was used can be downloaded from Shujun Li website that doesn’t take up too much hard drive space and downloads very quickly compared to the DEMs on the USGS website. A grid cell size of 1000 m x 1000 m was used in the term project DEM.
ArcET takes the information provided in the DEM to interpolate (using the Elevational Detrended Kriging for interpolation) temperature and vapor pressure for a whole region. Temperature, in most cases, is the driving force behind evapotranspiration, therefore is extremely important to get temperatures that are accurate at non-weather station locations. This is done by using an observed temperature at the closest weather stations and is interpolated based off of distance away from the weather station, elevation, and latitude. The elevations are also used in the empirical equations as direct elevation input values and are used to find other parameters related to climate. The climate data sets mentioned in the second paragraph up, where used in conjunction with the DEM to calculate the reference evapotranspiration grid.
To determine the crop evapotranspiration, the crop
coefficients, types, and distributions are needed. The crop coefficients where gathered from
literature from the FAO 56 (Allen et al. 1998).
From the crop growing stage, crop coefficients of developmental, middle,
and late where determined from tables.
Taking the values form the tables allowed for the monthly crop
coefficients to be estimated and tabulated as a data base file (DBF). The crop distribution shape file showed the spatial
distribution and types of the crops in
Problems with ArcET (That I found and
worked around)
ArcET, as it is when its downloaded, is not fully debugged and has a few programming and interface glitches. Despite this fact, ArcET will generate an extremely useful grid of reference evapotranspiration and crop coefficient map. The problem that I found in using ArcET was that the program would crash when attempting to convert the crop coefficient map into a raster grid. I think this problem develops from when the program doesn’t properly tabulate the crop coefficient values in the user interface. Therefore the crop coefficients are not joined with the corresponding crops on the shape file and when it goes to perform a raster multiplying calculation the reference evapotranspiration cannot be multiplied by the crop coefficient distribution. The purpose of the raster calculation is shown above in the bottom picture in Figure 1.
To work around this problem, the joining of tabular crop
coefficients data was needed to be done with the land cover/use shape
file. Using the join option, the
selected crop coefficients month (August in the analysis) was joined to the
shape file for the corresponding crops.
The newly updated land cover/use shape file was then converted to a
raster gird format using the conversion toolbar, by assigning the crop
coefficient as the spatially distributed value.
This gave the second raster in order to calculate the crop
evapotranspiration. The raster
calculator was used to multiply the reference evapotranspiration with the crop
coefficient to generate the crop evapotranspiration map. A table was generated with the crop evapotranspiration
map showing all the crop evapotranspiration within the area of interest so that
the results can be analyzed.
Project Application of ArcET Extension
The ArcET interface gives the user six different methods to choose from to determine reference evapotranspiration. Two of the empirical equations are identical, they just reference a different length of grass (tall and short), and therefore only five equations are shown in the results section. The first step is to download the extension and get the toolbar situated on the ArcGIS interface. From personal experience, it took three tries until a functional ArcET toolbar was downloaded. The first step to using ArcET is to input the needed data inputs under the general settings in the toolbar shown in Figure 3.
Figure 3: Shows the general settings interface of ArcET.
Under the general settings, there are five tabs. Under the first tab, the location of the gathered information is input, along with the data units as shown in the bottom of the Basic Data tab. The second tab is labeled “Methods for Calculating Reference ET” and allows for the user to select from six different empirical equation methods as listed previously. The second tab allows the user to insert any assumptions like wind speed and other parameters to calculate solar radiation and dew point temperature. The third tab gives the user the option to use interpolated rain data from the weather stations to use a PRISM precipitation grid that can be downloaded from (http://www.ocs.orst.edu/prism/ state_products/maps.phtml?id=UT) for the state of Utah. The fourth tab allows for the data output to be routed to the desired location. The fifth tab gives the user the option to show the percent of calculations complete when the reference evapotranspiration is being computed. The next step is to hit apply in the general settings and then go back into the ArcET toolbar dropdown menu and go into the evapotranspiration estimation, and then go the calculate reference evapotranspiration.
It will seem that the computer freezes up and the task manager indicates that the program is not responding, but it takes about eight minutes to perform the calculations to get the crop evapotranspiration map, just be patient, they will be generated. The steps outlined in this section, project application of ArcET extension, illustrates the entire use of ArcET in the term project, meaning it was only used to generate the crop reference raster grid.
Results of Reference Evapotranspiration
Each method’s reference evapotranspiration is displayed below for each method because all of the maps differ from one another, while the crop distribution and coefficients are the same raster grid to calculate the crop evapotranspiration. The reference evapotranspiration maps have the following classification in Figure 4 for showing the difference in evapotranspiration values so they can be easily compared by viewing the maps. It is important to note that the units for the following reference evapotranspiration maps are in mm/day based upon the monthly average of August. The last map is the crop coefficient values; there are not dimensions for these values.
Figure 4: Shows scale for
reference evapotranspiration
for the five following maps.
FAO 56 Penman-Monteith:
Min: 4.51 mm/day
Max: 8.38 mm/day
Standardized ASCE
Penman-Monteith:
Min: 4.51 mm/day
Max: 8.38 mm/day
Hargreaves 1985:
Min: 3.81 mm/day
Max: 8.55 mm/day
SCS modified Blaney-Criddle:
Min: 1.72 mm/day
Max: 9.69 mm/day
Priestley-Taylor:
Min: 4.89 mm/day
Max: 7.34 mm/day
Crop Coefficients grid:
Units are dimensionless
Conclusion
The results from the following maps have reference evapotranspiration values that range from approximately 1.7 mm/day to 9.7 mm/day. The method that had the greatest variance in reference evapotranspiration values was the SCS modified Blaney-Criddle method. While the least amount of variance was from the Priestley-Taylor method. When the reference evapotranspiration was combined with the crop coefficients grid to make the crop evapotranspiration, the results were tabulated and analyzed. What really stands out is how the maximum, minimum, average, and standard deviation of the crop evapotranspiration is similar for the first three methods listed in Table 1.
Table 1: Shows the extreme, average, and standard deviation of the crop evapotranspiration values.
Hand calculations for selected areas gave similar results, usually within 0.5 mm/day. When comparing evapotranspiration values from the Utah Department of Water Resource’s (UDWR) website (http://www.conservewater.utah.gov/et/etsite/ default.asp?summary.htm) for certain areas, evapotranspiration values where comparable, usually within 0.5 mm/day. All the weather data was gathered from the UDWR website to do the hand calculations. Usually the results from ArcET maps and tabulations gave results within 0.5 mm/day of each other, but had extreme differences as high as 3 mm/day.
It is difficult to recommend what method should be used as a standard for calculating reference evapotranspiration from the results of this analysis, because some equations will work better in some climates and regions than others. For example, the Hargreaves 1985 is best used in arid to semi-arid regions.
A practical method that could be applied to all regions would be to create maps and tabulated values for the area of interest and then compare the results and look for similarities. In this analysis, there where three different methods that yielded extremely similar results as observed from Table 1 (FAO 56 Penmon – Monteith, Standardized ASCE Penman – Monteith, and Hargreaves 1985). In this case, I think that there is substantial data that would lean to use the crop evapotranspiration of the three similar results. In the case that a region doesn’t yeild similar results, I would recommend using the more conservative crop evapotranspiration, or refer to local guild lines on calculating evapotranspiration and follow their standards. In Figure 5 is the crop evapotranspiration map that was generated using the FAO Penmon – Monteith method.
Figure 5: Shows the crop evapotranspiration map that was generated
using the FAO Penmon – Monteith method.
Works Cited
Li, Shujun, Tarboton, David, Mckee, Mac. "An ArcGIS
Extension for Regional ET Estimation." ArcET.
18 April 2005.
Division of Water Resources. (2006). “Summary Weather Data and ET Site.” 08 December 2006. < http://www.conservewater.utah.gov/et/etsite/
default.asp?summary.htm>.
Allen, Richard G., and Pereira, Luis S. (1998). Crop evapotranspiration – Guildlines for computing crop water requirements – FAO Irrigation and drainage paper 56. (accessed 2006, December 08). < http://www.fao.org/docrep/X0490E/ X0490E00.htm>.