Assessing Effects of Data Limitations on Salinity Forecasting in Barataria Basin, Louisiana, with a Bayesian Analysis
Assessing Effects of Data Limitations on Salinity Forecasting in Barataria Basin, Louisiana, with a Bayesian Analysis
复制标题
利用贝叶斯分析评估数据限制对路易斯安那州巴拉塔里亚盆地盐度预测的影响
DOI：
10.2112/06-0723.1
复制
发表时间：
2007
期刊：
影响因子：
3.2
通讯作者：
R. Twilley
中科院分区：
文献类型：
--
作者：
E. Habib;W. Nuttle;V. Rivera‐Monroy;S. Gautam;Jing Wang;E. Meselhe;R. Twilley
可靠的盐度变化预报对于恢复和维持河口和沿海生态系统的自然资源至关重要。由于这类生态系统的物理复杂性，应评估与盐度预测相关的不确定性信息，并将其纳入管理和恢复决策。这项研究的目的是调查路易斯安那州巴拉塔里亚盆地由于可用于校准和应用物质平衡盐度模型的数据的限制而造成的盐度预测的不确定性。该盆地是一个以河口湿地为主的生态系统，位于密西西比河三角洲复合体的正西部。该流域一直在以每年近23km2的速度流失湿地。应用基于贝叶斯的方法研究了数据相关的不确定性对模型参数的反演和随后的模型预测的影响。我们的重点是盐度校准数据的抽样和复盖率有限以及流域内稀疏的雨量计数据造成的不确定性。结果表明，由于数据的限制，模型参数的辨识存在很大的不确定性，导致模型结果中存在较大的系统误差和随机性误差。最显著的影响与缺乏关于降雨量的准确信息有关，降雨量是该盆地的主要淡水来源。这项研究的方法和结果可用来确定在监测复杂河口系统方面的必要改进，以减少预报的不确定性，并使管理人员在规划恢复沿海资源方面更加准确。
Abstract Reliable forecasts of salinity changes are essential for restoring and sustaining natural resources of estuaries and coastal ecosystems. Because of the physical complexity of such ecosystems, information on uncertainty associated with salinity forecasts should be assessed and incorporated into management and restoration decisions. The objective of this study was to investigate uncertainty in salinity forecasts imposed by limitations on data available to calibrate and apply a mass balance salinity model in the Barataria basin, Louisiana. The basin is an estuarine wetland-dominated ecosystem located directly west of the Mississippi Delta complex. The basin has been experiencing significant losses of wetland at a rate of nearly 23 km2/y. A Bayesian-based methodology was applied to study the effect of data-related uncertainty on both the retrieval of model parameters and the subsequent model predictions. We focused on uncertainty caused by limited sampling and coverage of salinity calibration data and by sparse rain gauge data within the basin. The results indicated that data limitations lead to significant uncertainty in the identification of model parameters, causing moderate to large systematic and random errors in model results. The most significant effect was related to lack of accurate information on rainfall, a major source of fresh water in the basin. The approach and results of this study can be used to identify necessary improvements in monitoring of complex estuarine systems that can decrease forecast uncertainty and allow managers greater accuracy in planning restoration of coastal resources.