Surface & Groundwater Water Demand by Local Drainage Areas (LDA's)

Dr. Jim Hook
University of Georgia, Tifton, GA

Several drainage areas (LDA's - black outline) and stream outlet nodes (green plus symbol) for a multi-county area of Southwest Georgia.
Statewide view of LDA's and Nodes
EPD hydrologists have the responsibility to estimate impacts of potential surface water withdrawals on flow in streams and levels in lakes and serervoirs. As part of this effort, EPD hydrologists have broken the state's surface water into multiple-stream watersheds that have defined outflow points. At many of these points (also called nodes) there is a stream gaging station that allows verification of modeled flow with flow measured under past conditions. These locations will also allow future flow measurements to be compared when measured withdrawals from municipal, industrial, and agricultural users are taken into account. To facilitate these model predictions, EPD has asked for a breakdown of predicted agriculural water demand through 2050 for each local drainage area (LDA). Summaries are needed by month for years when rainfall and other conditions cause irrigation to occur at very low to very high use (10, 25, 50, 75, and 90th percentiles.)

How were these projections made?

Withdrawal quantities were computed for each LDA as the product of three values:
  • Projected irrigated area for a crop (acres),
  • Predicted monthly irrigation application depth (inches),
  • Proportion of irrigation water derived from a source (fraction).
The product, monthly withdrawals (acre-inches) by crop, was summed for the LDA. To be consistent with other water planning efforts, acre-inches per month was converted to million gallons per day (MGD) by converting to gallons/month and dividing by the days in the month.
Limitations of the approach used in LDA computation of withdrawals: The existing Agricultural Water Demand for 2011 to 2050 was computed primarily for multi-county Water Planning Regions. At EPD's request it was disaggregated to county levels to allow a more detailed plan to be computed. That was possible because most water use records on which it was based - including Agricultural Water Use Permitting, Cooperative Extension Surveys - Annual Farm Gate Survey and triennial Irrigation Surveys, and USDA Surveys - Census of Agriculture and Farm and Ranch Survey - were available on a county level. Even historical weather records that were used in the model estimation of future irrigation application depths for each crop type were available for most counties in Georgia. Those 57 years of records were assumed to be reasonable proxies for future weather scenarios for 2011 through 2050.

Disaggregating the Agricultural Water Demand by hydrologic boundaries (LDA's) that extend across one or more counties create several challenges and demand new assumptions. These assumptions and the way the challenges were met are listed below:
  • Baseline irrigated area: The baseline for existing irrigated area in the AWD was assumed equal to the GIS-mapped field areas evident in 2005 through 2007 aerial imagery, Soil and Water Conservation Commission field mapping, and EPD mapping of farmer-supplied information concerning their permitted withdrawals (Procedures outlined previously). Fields that overlapped county borders were assigned to a single county based on county of primary area, as expressed by the polyXY (weighted center) of the mapped field shape. Because irrigated areas are known by GIS map locations, the same procedure was used to assign fields to LDA's. The fields retained their original county label, as well, for use in estimating crops, irrigation depths, and sources for LDA's, as explained below.

    After the initial Ag Water Demands, released July, 2009, were presented and analyzed, planners and agricultural water users requested more detail about minor use areas, and EPD hydrologists requested summaries to be made by Local Drainage Areas (the analysis provided on this page.). Other requests resulted in separate analyses. These included animal water use (Livestock and poultry water use) and water use by agriculturally permitted golf courses (golf course water use). Updates followed public review of AWD that noted omissions of certain non-permitted withdrawals used in irrigation, incorrectly labeled use or water source for mapped field areas, and omitted fields that could be identified through reexamination of 2005 to 2007 imagery or through recently or updated location of wells or pumps from the EPD Agricultural Permitting Unit. After the initial LDA analysis, the green industry requested a more detailed look at it's water use area and asked that even unpermitted uses in small nurseries be considered as well as that annual irrigation amounts more closely reflect irrigation as practiced by its major producers. In all, about 1500 additional fields and greenhouses were added, recategorized, or updated through this effort, through April 16, 2010. Updated and corrected field shapes were labeled with the word "POST" in the GIS data table.

  • Water sources within LDA's: Original AWD projections used a combination of mapped data with known water sources and county-based averages for irrigated fields with unknown water sources. Known sources are noted for each irrigated field as a percent ground water and percent surface water. For well-only supplies, field source is labeled 100% GW. For surface pump-only supplies, field source is labeled 100% SW. For those with known well-to-pond sources, field source is labeled 70% GW and 30% SW. Known sources included SWCC meter sites, EPD permitted and mapped sites, and fields mapped from aerial images when the field area is coincident to or over known pump sites. Unknown source sites were those mapped from aerial images when no permit is known for the area or when no obvious source assignment can be made. For unknown source fields, field source is labeled x% GW and y% SW, where x and y are the proportions of that irrigated area-weighted sources for known fields in the county.

    Assignment of irrigation water sources to the LDA's in the April, 2010, update ideltically as with counties in WPR's, i.e. for irrigated fields with known water sounces, those sources remained attached to those specific fields in LDA's. For fields with unknown water sources, assignment of probable sources was made based on the averages observed for known sources in the LDA in which the field is located. See data set - Water Sources Summarized by LDA, below.

    For fields that had permitted groundwater sources (or permitted well-to-pond sources), the aquifer assignment made by EPD was added. Those assignments were made based on information provided to EPD by the permit applicant or by the well driller. For fields with multiple wells used to supply or backup the irrigation, a single source assignment was made. In those cases, if EPD used multiple aquifer designations, they were used, otherwise the primary source was used. The total groundwater-supplied irrigated acres were divided among the various aquifers and a percentage by source, including "unknown" within each LDA was computed. Then the acres in the "unknown" source fields was apportioned to the known sources in proportion to their acres by source for the known aquifers within the LDA. Both summaries are provided. See data set - Water Sources Summarized by Aquifers within LDA's, below.
Dataset - Irrigated Fields w/Sources in LDA's
Excel Spreadsheet - Updated through 4/16/2010
Dataset - Water Sources Summarized by LDA's
Excel Spreadsheet - Updated through 4/16/2010
Dataset - Water Sources Summarized by Aquifers within LDA's
Excel Spreadsheet - Updated through 4/16/2010

  • Baseline crops in mapped fields: Most mapped fields in the baseline were labeled by the indefinite crop "rotation" label. This meant that the crop or crops present in any given year was subject to the farmers choice of economically viable production systems. In the baseline, we apportioned this indefinite area among the irrigated rotation crops reported county-by-county in the 2008 CES Irrigation Survey. These included corn, cotton, peanut, tobacco, soybean, and sprinkler irrigated vegetables. The county-by-county data assured that we would not apportion part of the rotation area to, say, peanuts, if the survey showed that irrigated peanuts were not grown in that county. It was assumed, though that any particular field could have any of that county's rotation crops present in any given year. Corn, for example was not assigned to one of those fields for the 40-year projection cycle. The exception to this was for fields identified as "permanent" orchards, i.e. pecan, peach, grape, apple, and berry, or as "permanent" green industry areas, i.e. nurseries, sod farms, and greenhouses. We understood that the term permanent may not apply to most of these crops over a 40-year cycle, but we assumed that irrigated acreage would remain in the locations where it existed and growth (or shrinkage) of that acreage would be proportional to the county's baseline orchard and green industry acreage.

    Assignment of crops to the LDA baseline area required an extrapolation/interpolation process using county data. First, for fixed (orchard) crops and green industry crops, use assignment in the LDA was made based on that field's original, mapped, use label. Produce and rotation areas, on the other hand, are expected to change crops each year. For these we assigned crops based on the proportion of a county's total irrigated area within an LDA and the county's rotation crop use. For example, if irrigation in a drainage area was made up of 10,000 acres county from A, 5,000 acres from county B, and 5,000 acres from county C, corn acreage in the LDA would equal to corn % in county A X 10,000 plus corn % in county B X 5,000 plus corn % in county C X 5,000, with the sum divided by 20,000.

Dataset - Baseline Irrigated Crops in 2008 by LDA's
Excel Spreadsheet - Updated through 4/16/2010 - Updated through 4/16/2010

  • Irrigation depths in LDA's: Irrigation depths were computed for each commodity within each county. Irrigation depth predictions were made in daily time using tha county's three dominant soil types, planting dates, and 57 year weather scenario. They were summed by months within year, and monthly data was ranked over years to obtain the 10, 25, 50, 75, and 90% probability levels for irrigation (Procedures outlined previously).

    Assignment of irrigation depth (probabilities) to the LDA's required an extrapolation/interpolation process using the data derived for the county. For each commodity an area-weighted value was calculated for each percentile and month, based on percent area of the county within the LDA. If a commodity was absent from one or more counties in an LDA, the depth value would be derived from remaining counties that had that crop present.

Dataset - Irrigation Depths by Crop, Month, & Probability within LDA's
Excel Spreadsheet - Updated through 4/16/2010

  • Monthly and Annual withdrawals for Baseline year and projection years 2011, 2020, 2030, 2040 & 2050: Irrigation withdrawals were calculated by month (in MG) for very wet to very dry conditions for fields (or portions of them) supplied from surface water sources in the LDA using the equation above. Note all surface water withdrawals were reduced from 100% of water needs by 30%, based on observation during 1999 to 2004 (Ag Water Pumping) that surface water users only applied 70% as much as GW users for the same crops and conditions. This assumption may not apply for users with adequate, year-round supplies in larger reservoirs on their property. Note most surface water withdrawals in the Coastal Plain LDAs are from farm pond reservoirs rather than flowing streams. Most of these ponds are located in ephemeral drainageways (personal communication, J.E. Hook, 2009).

  • Future crop irrigated areas were projected on a statewide basis by economists. Growth rates (+/-) were computed annually through 2050 for major commodities - corn, cotton, peanut, and soybean - as well as for nuts and fruit (orchards), fresh market vegetables, and process vegetables (produce). While individual rates were computed by commodity for those crops, averages over those were used to project future growth rates for crops for which projections could not be made - green industry crops, tobacco. These statewide growth rates were then disaggregated to county level by assuming each county with a projected commodity would partake in that growth in proportion to their baseline acres of that crop. Since baseline crop acres by LDA was computed in the previous step, this disaggregation used the same assumptions and assigned growth to LDA in proportion to statewide projections for each commodity.

Dataset - Baseline & Projected Monthly Water Withdrawals for Irrigation by Probability within LDA's
Excel Spreadsheet - Updated through 4/16/2010
Dataset - Baseline & Projected Annual Water Withdrawals for Irrigation by Probability within LDA's
Excel Spreadsheet - Updated through 4/16/2010
Cautions for data interpretation:
Withdrawals are given in Million Gallons. To compute daily withdrawal rates (MGD, divide those values by days in the summary period.

Many local drainage areas defined by nodes on the Savannah River that borders South Carolina and the Chattahoochee River where it borders Alabama include substantial areas of those states. In this analysis, only irrigation that occurs in Georgia is considered. Flow at the affected nodes, however, is influenced by withdrawals for agricultural irrigation, municipal water supplies, industry, and hydropower originating in the LDA's but outside of Georgia. A similar situation occurs with a few nodes situated within Tennessee, North Carolina, and Florida. Part of the withdrawals affecting those nodes may be from activities occurring outside of the State of Georgia.

Last updated 4/16/2010 James E. Hook
Links reconfirmed 11/7/2012 JEH