Numerical Estimation of the future sustainable Groundwater Yield in the Kok River Basin, Northern Thailand

Transcription

1 Numerical Estimation of the future sustainable Groundwater Yield in the Kok River Basin, Northern Thailand PHATCHARASAK ARLAI 1, MANFRED KOCH 2 and ARUN LUKJAN 1 1 Research Center of Water Resources and Disaster Mitigation Management, Nakhon Pathom Rajabhat University, Thailand riverine_eng@yahoo.com 2 Department of Geohydraulics and Engineering Hydrology, University of Kassel, Germany Abstract Kok river basin in the Golden Triangle delta in northern Thailand is becoming a major international trade hub in the region. Going hand in hand with the expected future economic growth in this area will be adverse environmental stress on the water resources in the Kok river basin. Although groundwater in the region is still abundant now, there is increasing concern among Thai authorities that future over-pumping may deplete parts of the aquifers there. For that reason a groundwater sustainability study was initiated in order to quantify the future sustainable extraction rates for the various aquifers underlying the Kok river basin. Using a calibrated 3D groundwater flow model sustainable extraction yields - defined as the maximal total pumping rate that ensures that piezometric heads in an aquifer do not fall below 20 meters from the land surface in the next 20 years have been determined for the four aquifers within the larger Kok river basin. The results indicate that there is still large room for near-future groundwater development in the study region. Key Words groundwater modeling, sustainable yield, Kok river basin, Thailand 1. INTRODUCTION The Kok river basin situated in the Golden Triangle delta in northern Thailand is becoming to serve as an important regional trade hub for Yunan Province in China, Myanmar, Laos P.D.R and Thailand. As a consequence, the Kok river basin is expected to experience major economic growth in the near future which, despite its positive social impacts, should exert some environmental stress on the natural resources of the region, namely, water resources. Imminent climate change across the region as a whole may further exacerbate these adverse effects. Although groundwater in the Kok river basin is still abundant at present time and is supposed to support the water needed to sustain the envisioned future economic growth, no thorough investigation with regard to the quantity (yield) and the future sustainability of the groundwater in the basin exists up-to-date. For this reason the Thai Department of Groundwater Resources (DGR) recently initiated a comprehensive study of the hydrogeology and groundwater resources of the Kok basin aquifers, with the objective to estimate the future sustainable groundwater yield. Groundwater modeling is one indispensable tool for achieving this goal. Here we report on first results of the application of the 3D groundwater flow model MODFLOW (Harbaugh and McDonald, 1996) to the Viang Pa Pao, Mae Suai, Upper Chiang Rai and the Mae Sai aquifers which make up most of the Kok river basin (see Fig. 1). Focus of this, yet preliminary, modeling study is then the estimation of the future sustainable groundwater yield for these aquifers, in order to plan well ahead for future sustainable groundwater development in the basin. 2. SET-UP OF THE CONCEPTUAL AND NUMERICAL GROUNDWATER MODEL 2.1 Conceptual models for the various aquifer systems in the Kok river basin The conceptual 3D- groundwater model represents the overall hydrogeological and physical characteristics of the aquifer system. That is, the conceptual model is set up based on the geological data, geological description, hydrogeology, hydrology, topography, groundwater extraction, soil conditions and land uses, some of which has only been recently collected (DGR, 2009). Four independent MODFLOW models have been set up for the four aquifer systems (Viang Pa Pao, Mae Suai, Upper Chiang Rai and Mae Sai) that underlie the Kok river basin (Fig. 1). Fig. 2 illustrates the conceptual models, namely, number of aquifer layers modeled, boundaries, recharge and discharge (pumping) zones, and river courses that are simulated by means of the corresponding

2 Fig.1 Land uses map (left panel) and geological map (right panel), modified from DMR (2005) packages of the MODFLOW model (Harbaugh and McDonald, 1996). Fig. 2 shows that the overall conceptual layouts for the four aquifer systems are very similar, i.e. no-flow boundaries underneath the ridges of the mountains encompassing a sub-basin; groundwater recharge areas though the surficial alluvial terraces at the hill-slopes long the Mae Nam Lao and Kok rivers, the latter being included into the model by means of MODFLOW s river package. Further details of the conceptual model for the Viang Papao sub-basin aquifer are provided in Arlai et al. (2010). One peculiarity occurs at the northern boundary of the Mae Sai aquifer system which is specified as a general head boundary, since there is a hydraulic connection between the Thai and the Burmanese sections of this sub-basin. It should also be noted that, based on the hydrogeological information (e.g. Arlai et al., 2010) all layers of the four aquifer systems (see Fig. 2) are in fact modeled as a sequence of aquifers and aquitards by specifying appropriate values for the hydraulic conductivity (see Section 3). 2.3 Groundwater flow model set-up and optimization of the model grid Following the build-up of the conceptual models of the four aquifer system, the 3D finite difference (FD) MODFLOW groundwater flow model was set up. with the number of layers and the boundary and external forcing conditions as illustrated in Fig. 2. Prior to the detailed calibration of the model (Section 3), the size of the optimal horizontal grid, i.e. the optimal number of horizontal grid points was determined for each of the aquifer systems. Doing so not only avoids unnecessary small grid sizes, i.e. excessive simulation times, but also discourages the use of very different widths and lengths of the elements of FD-grid which should be prevented in order to reduce numerical discretization errors (Domenico and Schwarz, 1998). The FD-grid was then optimized by repeated steady-state calibrations using coarse estimations of the important hydrological parameters that enter

3 (a) Conceptual Model of Viang Pa Pao Basin: 1 unconfined, 6 confined aquifer layers (b) Conceptual Model of Mae Suai Basin 1 unconfined, 7 confined aquifer layers (c) Model of upper Chiang Rai Basin: 1 unconfined, 1 confined aquifer layers (d) Model of Mae Sai Basin 1 unconfined, 7 confined aquifer layers Fig. 2 Conceptual models of the Viang Pa Pao-, Mae Suai-, upper Chiang Rai- and Mae Sai aquifer systems with boundaries, recharge zones and number of aquifer layers indicated (a) RMSE (m) versus grid size for the Mae Sai aquifer system. An optimal grid size of about 400 m is obtained. (b) FD-grid of Mae Sai aquifer system Fig.3 Optimization of the model grid size (a) and optimal FD-grid (b) for the Mae Sai aquifer system

4 (a) Scattered plot between observed- and computed head in steady state calibration (b) Comparison of observed-versus Inactive Cell computed head in transient calibration Fig.4 Steady state- and transient calibration of the Mae Sai aquifers model the model (Section 3) and checking the root mean squared error (RMS) between the observed and modeled hydraulic heads. Based on these initial model tests, grid sizes of 400x400, 500x500, 500x500 and 400x400 m 2 were obtained for the Viang Pa Pao-, Mae Saui-, upper Chiang Rai- and Mae Sai aquifers. Results of this grid optimization are shown for the Mae Sai aquifer in Fig STEADY-STATE AND TRANSIENT MODEL CALIBRATIONS Groundwater model calibration is usually carried out to ensure that the model can reasonably well mimic the groundwater flow system, namely, fit the observed piezometric heads. The latter were measured at 48 monitoring wells that were installed by DGR during The measurement period available in the present study is from January to June, The calibration was done in both steady and transient state. The horizontal hydraulic conductivity K xy in the various aquifer layers and the groundwater recharge W turned out to be the most sensitive calibration parameters and were adjusted in an iterative manner during the calibration process. For the Mae Sai aquifer systems some results of the calibrations in both steady-state and transient modes are shown in Fig. 4. Further details concerning the calibration for the Viang Pa Pao - aquifer system can be found in Arlai et al. (2010). 4. REGIONAL AND SUB-DISTRICTAL SUSTAINABLE GROUNDWATER YIELDS There have been many, often different, definitions of the concept of sustainable yield (SY) (cf. Alley et al., 1998; Alley and Leake, 2004; Maimone, 2004; Kalf and Woolley, 2005 and Arlai et al., 2006; 2007), which depend on the local aquifer system as well as on the environmental and political constraints within the region of interest. In the present case the Thai DGR has defined the SY of the aquifers under question as the maximum total pumping rate above the current pumping rate that ensures that the average piezometric head in each layer does not fall below a distance of 20 meter from the land surface in the next 20 years. This definition of the SY has been introduced into the groundwater flow model, whereby each active cell of the FD-model is associated with a pumping well. In addition, to account for the decreasing aquifer thickness at the horizontal boundaries of the models underneath the hill-slopes (see Fig. 2), the SY were calculated for two zones for each aquifer, namely, a zone 1 where the aquifer thickness is smaller than 50 m, and a zone 2, covering the remainder and the majority of the basin area with aquifer thicknesses larger than 50 meters. The initial pumping in each active cell for each zone has been calculated from the averaged values of the current total pumping rate in the corresponding zone. Based on these estimations, the initial pumping rates were specified in Viang Pa Pao-, Mae Saui-, upper Chiang Rai- and Mae Sai aquifers as 0.24 and CMD, 1.42 and 1.84 CMD, 3.08 and CMD and 1.61 and 1.64 CMD in each active cell of pumping zones 1 and 2, respectively. The future SY s for the next 20 years under the

5 Fig.5 An example of zoning of sustainable yield of the 1 st and 2 nd zone (left) and application of the district-wise zone budget model (right) to the Mae Sai aquifer system named constraint were then investigated. Eventually, total SY- values of 168,219 CMD, 35,519 CMD, 25,891 CMD and 565,941 CMD were obtained for the Viang Pa Pao-, Mae Saui-, upper Chiang Raiand Mae Sai aquifer systems, respectively. From these results one may notice that the Mae Sai aquifer system is the most productive and the upper Chiang Rai the least productive one. Finally, using the zone budget module of MODFLOW, the SY s were computed separately for each district of the four aquifer basins. The results of these computations are listed in Table 1 for the various districts (Fig. 5) of the Mae Sai aquifer system. For this aquifer district zone 11 turns out to be the most productive one, with a developable pumping rate (= sustainable yield minus present-day pumping rate) of 1706 CMD/km 2, while the zone 1 is the least productive one with a developable pumping rate of only 521 CMD/km 2. For a detailed discussion of the corresponding results for the Viang Pa Pao- aquifer system the reader is referred to Arlai et al. (2010). Table 1 District-wise computation of the sustainable yields (SY) for the Mae Sai aquifer system. Also indicated is the developable pumpage (=sustainable yield minus present-day pumping rate) Zone Area Present Pumpage Pumpage under Yield Constraint Developable Pumpage (km 2 ) (CMD) (CMD) (CMD) (CMD/km 2 ) Total

6 4. CONCLUSIONS Sustainable future groundwater yields have been computed for four aquifer systems in the Kok river basin in northern Thailand which, due to accelerating economic development, may face significant water shortage problems in the near future, a situation that may be further exacerbated by imminent climate change in the region (Koch, 2008). A 3D groundwater flow model (MODFLOW) that takes into account the multi-layer structure of the various aquifers has been calibrated in steadystate and transient mode using recently measured piezometric heads in the aquifers. The sustainable yield has been defined as the maximum total pumping rate that ensures that the average piezometric head in each aquifer layer does not fall below 20 meter from the land surface in the next 20 years. Employing this constraint the transient MODFLOW computations result in total sustainable yields of 1.3, 12.36, 2.71 and m 3 /Rai (1600 m 2 )/day for the Upper Chiang Rai, Wiang Papao, Mae Suai and the Mae Sai aquifer, respectively. These yield variations reflect the different hydraulic conditions and hydrogeology of the four aquifer systems. Thus, the Mae Sai aquifer is the most productive one, since its hydraulic conductivity goes up to 190 m/day and it has a total thickness of 200 meters which resulting in rather high transmissivity values as well. Although the Upper Chiang Rai aquifer has also high horizontal hydraulic conductivities of up to 494 m/day, it is the least productive one, as its total thickness is just 44 meters. Using a zone budget module within the groundwater model, the sustainable yields have then been calculated also district-wise. Although still preliminary, the present modeling study should give policy makers a first tool at hand for future sustainable groundwater resources management in the Kok river basin. The next step of the study should be the incorporation of the effects of regional climate change on the sustainable yield estimation, as changing future seasonal rainfall pattern and temperatures will most likely affect groundwater recharge to the regional aquifers in an adverse way, putting extra stress on the groundwater system. REFERENCES Alley, W.M., T.E. Reilly and O.L. Franke (1999) Sustainability of Ground-Water Resources, U.S. Geological Survey Circular 1186, USGS, Reston, VI, US. Alley, W.M. and S.A. Leake (2004) The Journey from Safe Yield to Sustainability, Groundwater, 42, Anderson, M.P and W.W. Woessner (1992) Applied Groundwater Modeling: Simulation of Flow and Advective Transport, Elsevier, Amsterdam, The Netherlands. Arlai, P. (2007) Numerical Modeling of possible Saltwater Intrusion Mechanisms in the Multi-Layer Coastal Aquifer System of the Gulf of Thailand, Ph.D.Thesis, Kassel University, Germany. Arlai, P., M. Koch and A. Lukjan (2010) Modeling Investigation of the sustainable Groundwater ield for the Wiang Pao Aquifers System, Northern Thailand, Water Resources Manag. (submitted). Arlai, P., M. Koch and S. Koontanakulvong (2006) Modeling Flow and Transport for sustainable Yield Estimation of Groundwater Resources in the Bangkok Aquifer System, EGU, Vienna. Department of Groundwater Resources (2009) Groundwater Resources Assessment in the Kok River Basin, Bangkok, Thailand. Harbaugh, A.W. and M.G. McDonald (1996) Programmer's documentation for MODFLOW-96, an update to the U.S. Geological Survey modular finite-difference ground-water flow model: U.S. Geological Survey Open-File Report, , 220 pp. Kalf, F. R.P. and D.R. Woolley (2005) Applicability and methodology of determining sustainable yield in groundwater systems, Hydrogeol. Journal, 13, Koch, M., (2008) Challenges for future sustainable water resources management in the face of climate change, In: Proceedings of The 1 st NPRU Academic Conference, Nakhon Pathom University, Thailand, October 23-24, Maimone, M. (2004) Defining and Managing Sustainable Yield, Groundwater, 42,