Management Plan for Water Use to Improve a Wildlife Refuge Surrounded by an Agricultural Community in Southwest Puerto Rico
Eric Harmsen’s Sections
Justification
The Laguna Cartagena is in direct connection with the groundwater system. A portion of the laguna water is derived from the alluvial material deposited near the surface, and a portion from the upward flow of water through the fractured limestone aquifer. Laguna water originating from groundwater recharge within upland agricultural land, may contain elevated nitrates and/or pesticides. Laguna water originating within the limestone aquifer may contain elevated dissolved solids. By using a groundwater flow model in combination with particle tracking technique it will be possible to predict the source of the Laguna Cartagena water quality.
Objective
- Characterize the surface and subsurface flow systems of the LajasValley, with emphasis on the hydrologic interrelationships between these systems and the Laguna Cartagena.
- Develop a numerical groundwater flow model for predicting groundwater levels and flow directions within the LajasValley.
Background
The LajasValley is located in the extreme southwest of Puerto Rico (Figure 1). The 130 square km valley is oriented in an east-west direction, 35.4 km in length and 6.4 km in width. Unconsolidated material fills the valley which is flanked on the north and south by volcanic and limestone rocks of Cretaceous age. Historically, the area has had problems due to water logging and high salinity of the groundwater and soils. Anderson (1977) presented a detailed study of the hydrologic resources of the LajasValley. A summary of the hydrologic characteristics of the LajasValley are presented below.
Figure 1. Location of Lajas, PR (Magaly Revera, 2003).
The valley is a closed desert basin similar to those found in the Southwest U.S. Fine grained alluvium exists throughout the middle of the valley, while coarser grained material is found in the alluvial fans along the foothills. Five soil associations exist within the valley including the Fraternidad-Aguirre-Cartagena, Fé-Guánica-Aguirre, Americus-Guayabo-Sosa, Guayama-Aguilita-Amelia, Descalabrado-Jacana-San German Associations. Figure 1 shows the areas associated with the five soil associations. Figure 1 also shows the general layout of the valley, including the location of the Laguna Cartagena and the network of irrigation and drainage channels.
Rainfall varies from around 45 inches per year in the north area to less than 30 inches per year along the southern coast. Evapotranspiration is between 30 in per year in the non-irrigated areas to over 50 inches per year in the irrigated areas. Aquifer recharge comes from intermittent stream that head in surrounding mountains. While some of the floodwaters reach the “playa,” most of this surface runoff enters the alluvial fans where it recharges the alluvial aquifer. Groundwater in the western LajasValley principally discharges to the Bahía Boquerón. In the eastern portion of the valley groundwater under artesian pressure, leaks upward through relatively impermeable soil, where it is lost to evapotranspiration from the soil or seeps to drainage canals.
Water table or unconfined aquifers occur generally within the alluvial fans and mountain areas along the north and south. Alluvial fans in the foot hills consist primarily of sand and gravel material. In the northern area, water withdrawn from alluvial fans is of high quality (i.e., low salinity). Alluvium filling the central portion of the valley is mainly silt and clay material interspersed with sand stringers. Water withdrawn from this material tends to be brackish. Aquifers also occur within the consolidated limestone, sandstone units. In La Plata basin in the northeastern part of the valley, wells tapping limestone overlain by less permeable alluvium are confined or artesian (i.e., the wells penetrating this unit are flowing wells). Confined aquifers also occur in the eastern central portion of the valley near the Ciénaga El Anegado-Laguna de Guánica. The volcanic rock which bounds the valley in the north and south are nearly impermeable.
In the Laguna Cartagena area the major aquifer is a buried limestone unit, is highly permeably and considered to be unconfined. Water levels in the vicinity of the Laguna Cartagena in March 1965 were approximately 3 m (above mean sea level), and water levels in the extreme eastern portion of the valley were around 9 m. Water levels in the foothills were around 17 m(Figure 3). Typical annual groundwater level fluctuations are on the order of 1 m. In March of 1986, groundwater elevations weresignificantly higher near the Laguna Cartagena (13 m), but were lower in the extreme western portion of the valley (2 m) (Figure 4). Figure 4 clearly shows that a groundwater divide corresponds with East-West Drainage Divide.
Figure 4. Groundwater elevation within the LajasValley alluvial aquifer, March 1965 (Anderson, 1977).
Figure 3. Groundwater elevation within the LajasValley alluvial aquifer, March 1986
(Graves, 1991).
As a part of the LajasValleystudy, Anderson (1977)developed a computer model to analyze the potential use of a series of pumping wells to reduce water logging in the vicinity of Laguna de Guánica. The transient computer model simulated flow only in the upper alluvial aquifer. From the report it was difficult to determine the size of the model; apparently it was limited to the area of interested and did not include the entire valley.
Materials and Methods
1. Review of existing literature
As part of Objective 1, the surface and subsurface flow systems of the LajasValley will be characterized, with emphasis on the hydrologic interrelationships between these systems and the Laguna Cartagena. Numerous studies have been conducted on the LajasValley, considering for example, agricultural drainage andsalinity, geology and water resources. A geographic information system (GIS) has recently been developed for the Lajas Valley (Perez Alegria, 2003), which includes information on soils, streams, areas of flooding, salt-effected areas, farms, roads, and the irrigation and drainage canals. The GIS will be of great value in the development of the proposed numerical groundwater flow model.
2. Field Study
It is necessary to determine the current conditions within the valley with respect to groundwater levels, especially in the vicinity of the Laguna Caragena. As part of the scope of work, permission will be obtained for measuring water levels within the approximately fifty existing wells within the alluvial aquifer. Approximately eight wells will be constructed in the alluvial aquifer where data is lacking. All new wells will be installed in compliance with EPA standards for chemical sampling and for measuring piezometric head. A groundwater elevation map will be developed for the entire valley for comparison with existing historical groundwater elevation maps. Water levels will also be obtained within the deeper bedrock aquifer to estimate vertical hydraulic gradients. All wells will be sampled for dissolved ions for the purpose of assessing the distribution of salinity within the groundwater systems. Nitrates will be evaluated to assess loading from agricultural land. A limited number of samples will be analyzed for pesticides. [CARLOS, CAN YOU DISCUSS METHODS TO BE USED FOR CHEMICAL ANALYSIS]
Groundwater Model
As an integral part of the model development a conceptual model of the hydrologic environment will be developed. A conceptual model represents our understanding of the system both qualitatively and quantitatively. The conceptual model includes the following components: topography and soils, climate, geology, hydrologic properties, aquifer properties (e.g., thicknesses, presence of confining layers, hydraulic conductivity, storitivity), aquifer recharge
aquifer discharge, groundwater levels (historical trends), tidal effects, groundwater flow directions, distribution of pumping wells and pumping rates, saltwater intrusion
soil and groundwater contamination, aquifer transport properties.
model codes
model design and development
source characteristics
initial conditions
boundary conditions
model calibration and sensitivity analysis
groundwater flow and solute transport models
model code selection
model design and development
spatial characteristics
grid design
initial conditions
boundary conditions
steady-state
temporal
model calibration, verification and sensitivity analysis
A groundwater model will be constructed using the numerical model Modular Three-Dimensional Ground-Water Flow Model (MODFLOW) developed by McDonald and Harbaugh (1984). Because of its ability to simulate a wide variety of systems, its extensive publicly available documentation, and its rigorous USGS peer review, MODFLOW has become the worldwide standard ground-water flow model. MODFLOW is used to simulate systems for water supply, containment remediation and mine dewatering. When properly applied, MODFLOW is the recognized standard model used by courts, regulatory agencies, universities, consultants and industry.
Where sufficient data exist, steady-state and transient model calibrations will be performed. Many of the areas being modeled will be based on data from previously calibrated groundwater flow models. However, for various reasons, it may be necessary to configure the larger regional-scale model differently from the smaller local-scale models (e.g., because different grid spacing may be used), and these differences may have an effect on the simulated groundwater levels.
The groundwater flow model will initially be calibrated for long-term average steady-state conditions. Calibration will be achieved by adjusting aquifer properties within reasonable limits in order to match observed average groundwater levels and discharge rates. Discharges will include base flow to rivers and discharges to the ocean. These data will be obtained from published reports. We will perform the MODFLOW calibration with the assistance of a commercially available nonlinear optimization program such as PEST (Doherty, 1994). In addition, a one-year transient model calibration will be performed in aquifers where synoptic groundwater level and discharge data exist.
Model Validation
We propose validating the model in two ways.
- Compare model estimates with ground-based historical data; and
- Compare the model-estimated island-wide water balance with a water balance obtained from ground-based and remotely-sensed data.
Validation Step 1
The data used for Validation Step 1 will be of the same form as was used in the model calibrations (i.e., from data collection stations), except that the data will be selected from different years. For example, if the transient calibration data for the groundwater flow model were from 1994, the validation data set would be from some other year, preferably a year with significantly different conditions (e.g., more wet or more dry).
Validation Step 2
The monthly water balance for the island, over a period of one year, will be calculated using the following simple equation:
DP = P + ET + RO + BF - S(1)
where DP is deep percolation or aquifer recharge, P is precipitation, ET is evapotranspiration, RO is surface runoff, BF is river base flow, and S is change in moisture storage. Each component of equation 1 is a function of space and time. On average S is negligible for long periods (e.g., one year), however, it will be important for shorter periods (e.g., one month).
The components on the right-hand-side of equation 1 will be estimated using ground-based and remotely-sensed data. Soil moisture will be estimated using the coupled hydrologic/radiobrightness model (Laymon et al., 2002) with data from the Advanced Microwave Scanning Radiometer-EOS (AMSR-E). Daily soil moisture content will also be estimated using a simplified water budget approach. The GIS-based water budget procedure is as follows:
- Infiltration will be estimated by subtracting surface runoff from rainfall. Runoff will be estimated using the curve number (CN) approach. Soils data (e.g., CN and soil moisture holding capacity) currently exist in GIS form for PR.
- If water initially within the soil profile plus the infiltrating water does not exceed the soil water holding capacity, then soil moisture content is equal to the initial volume plus infiltration.
- If water initially within the soil profile plus the infiltrating water exceeds the soil water holding capacity, then the excess water will be considered percolation and the soil moisture content will be adjusted to the value of the soil moisture holding capacity.
Remotely-sensed evapotranspiration will be obtained from the Aqua/MODIS system. Evapotranspiration will also be determined (within a GIS) using simplified procedures for estimating average monthly climate data for PR described by Harmsen et al., 2002.
Microwave, Surface and Precipitation Products (MSPPS) suite of products, which includes rain rate, land surface temperature, and land-surface emissivity will be obtained from NOAA/NESDIS. These are proven hydrological data products produced from the NOAA polar orbiters and are updated globally every four hours. These products have coarse spatial resolution (16-48 km) and will be downscaled. The most useful NASA products will be those generated from MODIS data. MODIS data can now be obtained from both the Terra (AM) and Aqua (PM) satellites. They have good temporal (1-2 days) and spatial (1-km) resolution. Products include land-surface temperature, land-cover type, vegetation indices, and leaf area index.
Conduct simulation studies of the hydrological cycle on Puerto Rico.
A series of short-term simulations on the order of days and months will be conducted to determine the sources and sinks of the precipitation across the island and parameters that could influence the hydrological balance. The hydrological sources and sinks will be stratified into evapotranspiration, runoff, soil storage, aquifer recharge, precipitation from convective clouds, frontal systems, and easterly waves, etc. These simulations will be configured for the following scenarios:
The simulations included a scenario in which thirty-five closely spaced pumping wells were located north and east of the Laguna de Guánica. Each well was pumped at 50 gallons per minute (gpm).
REFERENCES
Anderson, H. R., 1977. Ground water in the LajasValley: U.S. Geological Survey Water-Resources Investigations Report 68-76, 45 p.
Doherty, J. 1994. PEST Model Independent Parameter Estimation. Watermark Company.
Graves, R. P. 1991. Ground-water resources in LajasValley, Puerto Rico. U.S. Geological Survey Water-Resources Investigation Report 89-4182. 55 p.
Magaly Rivera, 2003. Welcome to Puerto Rico.
McDonald, M. G. and A. W. Harbaugh. 1984. A Modular Three-Dimensional Finite Difference Ground-Water Flow Model. U.S. Geological Survey.