GroMoPo Metadata for Edwards Aquifer USGS model

In 2010, the U.S. Geological Survey, in cooperation with the San Antonio Water System, began a study to assess the brackish-water movement within the Edwards aquifer (more specifically the potential for brackish-water encroachment into wells near the interface between the freshwater and brackish-water transition zones, referred to in this report as the transition-zone interface) and effects on spring discharge at Comal and San Marcos Springs under drought conditions using a numerical model. The quantitative targets of this study are to predict the effects of higher-than-average groundwater withdrawals from wells and drought-of-record rainfall conditions of 1950-56 on (1) dissolved-solids concentration changes at production wells near the transition-zone interface, (2) total spring discharge at Comal and San Marcos Springs, and (3) the groundwater head (head) at Bexar County index well J-17. The predictions of interest, and the parameters implemented into the model, were evaluated to quantify their uncertainty so the results of the predictions could be presented in terms of a 95-percent credible interval. The model area covers the San Antonio and Barton Springs segments of the Edwards aquifer; the history-matching effort was focused on the San Antonio segment. A previously developed diffuse-flow model of the Edwards aquifer, which forms the basis for the model in this assessment, is primarily based on a conceptualization in which flow in the aquifer is predominately through a network of numerous small fractures and openings. Primary updates to this model include an extension of the active area downdip, a conversion to an 8-layer SEAWAT variable-density flow and transport model to simulate dissolved-solids concentration effects on water density, history matching to 1999-2009 conditions, and parameter estimation in a highly parameterized context using automated methods in PEST (a model-independent Parameter ESTimation code). In addition to the best-fit parameter values derived from history matching, the uncertainty of model parameters was also estimated by using linear uncertainty analysis. Comparison of prior (before history matching) and posterior (after history matching) variances of parameters indicate that the information within the observation dataset used for history matching informs many parameters. The concentration threshold parameters were well-informed by the observation dataset as their posterior distributions were much narrower than their prior distributions. The transition-zone scaling parameters of hydraulic conductivity, effective porosity, and specific storage were all informed by the observation dataset, as evidenced by the difference between the prior and posterior variances. Saline-zone scaling parameters, alternatively, were not informed by the observation dataset for effective porosity and specific storage. Resulting posterior drier-month, wetter-month, and annual recharge multiplier parameter variances are important to understanding how well recharge is estimated and implemented within the model. The shifts of the posterior distributions left and right indicate that there were zones where less or more water was needed in the model. The widths of the distributions were not decreased substantially, indicating that many of the best-fit recharge parameters are not statistically different from the initial values specified in the history-matching effort. Recharge from rainfall is the driving force behind groundwater flow and heads in the aquifer; therefore, an increase in understanding of this process would benefit model development by potentially decreasing the uncertainty of this parameter. The history-matching effort was most helpful in informing the parameters in the model that control discharge at springs, namely, the spring orifice (drain) altitude and drain conductance parameters for each spring. The uncertainty assessment of the predictive model (a hypothetical recurrence of 1950-56 drought conditions and higher-than-average groundwater withdrawals from wells) provided insights into the potential effects of these conditions on dissolved-solids concentration changes at production wells near the transition-zone interface, discharges at Comal and San Marcos Springs, and heads at Bexar County index well J-17. Results at the 25 production wells near the transition-zone interface indicate that the uncertainty of model input parameters based on expert knowledge yielded an upper bound of the 95-percent credible interval of dissolved-solids concentrations that exceeds the secondary drinking water standards of 1,000 milligrams per liter (mg/L) of the Texas Commission on Environmental Quality (TCEQ) for many wells. However, the history-matching process provided key information to inform prediction-sensitive model parameters and therefore, contributed to a substantial decrease of the upper bound of the 95-percent credible interval to below the secondary drinking water standards. Reductions in dissolved-solids concentration changes were on the order of 400 mg/L to 1,300 mg/L. The reduction in uncertainty in regards to this prediction implies that this prediction of dissolved-solids concentration change can be made with some certainty using this current model and that those parameters that control this prediction are informed by the observation dataset. Even though predictive uncertainty was reduced for this prediction, dissolved-solids concentration changes were still greater than zero, indicating a minimal increase in concentration at these 25 production wells during the 7-year simulation period is likely. However, this minimal concentration increase indicates a small potential for movement of the brackish-water transition zone near these wells during the 7-year simulation period of drought-ofrecord (1950-56) rainfall conditions with higher-than-average groundwater withdrawals by wells. Predictive results of total spring discharge during the 7-year period, as well as head predictions at Bexar County index well J-17, were much different than the dissolved-solids concentration change results at the production wells. These upper bounds are an order of magnitude larger than the actual prediction which implies that (1) the predictions of total spring discharge at Comal and San Marcos Springs and head at Bexar County index well J-17 made with this model are not reliable, and (2) parameters that control these predictions are not informed well by the observation dataset during historymatching, even though the history-matching process yielded parameters to reproduce spring discharges and heads at these locations during the history-matching period. Furthermore, because spring discharges at these two springs and heads at Bexar County index well J-17 represent more of a cumulative effect of upstream conditions over a larger distance (and longer time), many more parameters (with their own uncertainties) are potentially controlling these predictions than the prediction of dissolved-solids concentration change at the prediction wells, and therefore contributing to a large posterior uncertainty.2010年，美国地质调查局（U.S. Geological Survey）与圣安东尼奥供水系统（San Antonio Water System）合作启动一项研究，旨在借助数值模型（numerical model）评估爱德华兹含水层（Edwards aquifer）内的半咸水（brackish-water）运移情况——具体而言，即淡水-半咸水过渡带（freshwater and brackish-water transition zones）附近的生产井（production wells）遭遇半咸水入侵的潜在风险，以及干旱条件下科马尔泉（Comal Springs）与圣马科斯泉（San Marcos Springs）的泉水流量（spring discharge）变化影响，本报告中将该过渡带的界面称为过渡带界面（transition-zone interface）。本研究的量化目标为：预测高于平均水平的地下水开采量（groundwater withdrawals），以及1950-1956年的记录级干旱降雨事件（drought-of-record rainfall conditions），对以下三方面的影响：(1) 过渡带界面附近生产井的溶解固体浓度（dissolved-solids concentration）变化；(2) 科马尔泉与圣马科斯泉的总泉水流量；(3) 贝尔克斯县基准井J-17的地下水水头（groundwater head，简称head）。本研究对目标预测及模型中引入的参数开展不确定性量化评估，最终以95%可信区间（95-percent credible interval）的形式呈现预测结果。 本模型覆盖爱德华兹含水层的圣安东尼奥段与巴顿泉段，历史拟合（history-matching）工作主要聚焦于圣安东尼奥段。本次评估所依托的爱德华兹含水层弥散流模型（diffuse-flow model），其核心概念框架为：含水层内的水流主要通过大量小型裂隙与孔隙网络运移。该模型的主要更新内容包括：向下延伸了有效模拟区域；转换为8层SEAWAT变密度流与输运模型（SEAWAT variable-density flow and transport model），以模拟溶解固体浓度对水体密度的影响；将拟合基准更新至1999-2009年的水文条件；在高度参数化的场景下，借助PEST（模型独立的参数估计代码，Parameter ESTimation code）的自动化方法完成参数估算（parameter estimation）。 除了从历史拟合中得到的最优拟合参数值（best-fit parameter values）外，本研究还通过线性不确定性分析（linear uncertainty analysis）估算了模型参数的不确定性。对比参数的先验分布（prior distribution）与后验分布（posterior distribution）的方差可知，历史拟合所用的观测数据集可为多数参数提供约束信息。浓度阈值参数的后验分布远窄于先验分布，说明观测数据集可有效约束该类参数。导水率（hydraulic conductivity）、有效孔隙度（effective porosity）、比储水率（specific storage）的过渡带缩放参数，同样可通过观测数据集得到约束，先验与后验方差的差异可佐证这一点。而咸水区缩放参数中，有效孔隙度与比储水率未被观测数据集约束。 补给乘数参数（recharge multiplier parameter）的后验方差（对应枯水月、丰水月与年度补给），对理解模型中补给量的估算与实施精度至关重要。后验分布的左右偏移表明，模型中部分区域需要调整补给水量。分布宽度未出现显著收窄，说明多数最优拟合补给参数与历史拟合阶段设定的初始值无统计学差异。降雨补给是含水层内地下水流与水头的核心驱动力，因此加深对该过程的理解，有望通过降低该类参数的不确定性助力模型优化。 历史拟合工作对约束控制泉水流量的模型参数最为有效，即各泉的泉口（排水口）高程与排水导纳（drain conductance）参数。本预测模型的不确定性评估（模拟1950-1956年干旱事件重现及高于平均水平的地下水开采），揭示了上述情景对以下方面的潜在影响：过渡带界面附近生产井的溶解固体浓度变化、科马尔泉与圣马科斯泉的泉水流量，以及贝尔克斯县基准井J-17的地下水水头。 针对过渡带附近的25口生产井的分析结果显示，基于专家知识的模型输入参数不确定性，使得多数生产井的溶解固体浓度95%可信区间上限超过了德克萨斯州环境质量委员会（Texas Commission on Environmental Quality，简称TCEQ）制定的次级饮用水标准（secondary drinking water standards）1000毫克/升（mg/L）。但历史拟合过程可为预测敏感型模型参数提供关键约束，因此将95%可信区间的上限显著降低至次级饮用水标准以下，降幅约为400 mg/L至1300 mg/L。该预测的不确定性降低表明，依托当前模型可对溶解固体浓度变化做出具有一定置信度的预测，且控制该预测的参数可通过观测数据集得到约束。尽管该预测的不确定性有所降低，但溶解固体浓度变化仍大于0，说明在本次7年模拟期内，这25口生产井的浓度存在小幅上升。而该小幅浓度上升意味着，在1950-1956年记录级干旱降雨事件叠加高于平均水平的地下水开采情景下，这些井附近的半咸水过渡带存在小幅运移的潜在可能。 7年模拟期内的总泉水流量预测结果，以及贝尔克斯县基准井J-17的水头预测结果，与生产井的溶解固体浓度变化结果差异显著。其可信区间上限较实际预测值高出一个数量级，这意味着两点：(1) 依托本模型得到的科马尔泉、圣马科斯泉总泉水流量及贝尔克斯县基准井J-17地下水水头的预测结果并不可靠；(2) 尽管历史拟合阶段得到的参数可重现该时段内的泉水流量与水头，但控制上述预测的参数在历史拟合过程中未得到观测数据集的有效约束。此外，由于这两处泉水的流量与基准井J-17的水头，更能反映更大范围、更长时间尺度上游水文条件的累积效应，相较于生产井的溶解固体浓度变化预测，控制该类预测的参数更多（且各参数均带有不确定性），因此最终导致后验不确定性显著增大。