Integrated hydrological modeling of the No-Name watershed, Medicine Bow Mountains, Wyoming (USA)

Abstract

Integrated modeling of headwater watersheds in mountain environments is often limited by the lack of hydrological characterization and monitoring data. For the No-Name watershed in the Medicine Bow Mountains in Wyoming (USA), this research integrates regional surface and subsurface hydrological and geophysical measurements to create three-dimensional integrated hydrological models with which interactions between surface water, soil water, and groundwater are elucidated. Data used to build and calibrate the integrated model include a digital elevation model (DEM), stream discharge at the outlet of the watershed, soil-moisture data, weather data, and geophysical surveys including seismic refraction, airborne resistivity, and nuclear magnetic resonance (NMR). Based on interpretation of geophysical measurements, subsurface hydrostratigraphy consists of a top unconsolidated layer, a middle layer of fractured granite and metamorphic bedrock, and a lower protolith. Given that both measurements and interpretations have uncertainty, a sensitivity analysis was carried out to evaluate conceptual model uncertainty, which suggests the following: (1) for predicting stream discharge at the No-Name outlet, the most influential parameters are the Manning coefficient, DEM, hydrostratigraphy and hydraulic conductivity, and land cover. Compared to a lower-resolution DEM, a LiDAR DEM can lead to more accurate predictions of the stream discharge and stream elevation profile. (2) For predicting soil moisture, the most influential parameters are hydrostratigraphy and the associated hydraulic conductivities and porosities. (3) Based on a calibration exercise, the likely values for subsurface hydraulic conductivity at No-Name are ~10^–5 m/s (the unconsolidated layer), ~10^–6 m/s (fractured bedrock), and ~10^–6 m/s or lower (protolith).

Résumé

La modélisation intégrée des bassins versants de tête dans des environnements de montagne est souvent limitée par le manque de caractérisation hydrogéologique et de données d’observation. Pour le bassin versant de No-Name, dans les Montagnes de Medicine Bow, au Wyoming (Etats-Unis d’Amérique), cette recherche intègre des mesures hydrologiques et géophysiques régionales de surface et de subsurface destinées à créer des modèles hydrologiques intégrés tridimensionnels, grâce auxquels les interactions entre l’eau de surface, l’eau du sol et l’eau souterraine sont élucidées. Les données utilisées pour construire et caler le modèle intégré comprennent: un modèle numérique de terrain (MNT), le débit du cours d’eau à l’exutoire du bassin versant, des données sur l’humidité du sol, des données météorologiques, et des levés géophysiques incluant la sismique réfraction, la résistivité aéroportée et la résonance magnétique nucléaire (RMN). D’après l’interprétation des mesures géophysiques, l’hydrostratigraphie de subsurface consiste en une couche supérieure non consolidée, une couche médiane faite d’un socle granitique et métamorphique fracturé et un protolithe inférieur. Etant donné que les mesures et les interprétations sont incertaines, une analyse de sensibilité a été réalisée pour évaluer l’incertitude du modèle conceptuel, laquelle suggère ce qui suit. (1) Pour prédire le débit d’écoulement à l’exutoire de NO-Name, les paramètres les plus influents sont le coefficient de Manning, le MNT, l’hydrostratigraphie, la conductivité hydraulique et l’occupation des sols. Par rapport au MNT de basse résolution, le MNT Lidar peut conduire à des prédictions plus fiables du débit et du profil d’élévation de l’écoulement. (2) Pour prédire l’humidité du sol, les paramètres les plus influents sont l’hydrostratigraphie et les conductivités et porosités hydrauliques associées. (3) Sur la base d’un exercice de calage, les valeurs vraisemblables de conductivité hydraulique de surface à No-Name sont de ~10^–5 m/s (couche non consolidée), ~10^–6 m/s (socle fracturé), et ~10^–6 m/s ou moins (protolithe).

Resumen

El modelado integrado de cuencas hidrográficas de cabecera en ambientes de montaña suele verse limitado por la falta de datos de caracterización hidrológica y de monitoreo. En el Caso de una cuenca situada en las montañas medicine Bow de Wyoming (EE.UU.), esta investigación integra mediciones hidrológicas y geofísicas regionales superficiales y subsuperficiales Para crear modelos hidrológicos integrados tridimensionales con los que dilucidar las interacciones entre el agua superficial, el agua del suelo y las aguas subterráneas. Los datos utilizados Para construir y calibrar el Modelo integrado incluyen: un Modelo digital de elevación (DEM), el caudal de los arroyos a la Salida de la Cuenca, datos sobre la humedad del suelo, datos meteorológicos y estudios geofísicos que incluyen la refracción sísmica, la resistividad aerotransportada y la resonancia magnética nuclear (NMR). Basándose en la interpretación de las mediciones geofísicas, la hidroestratigrafía del subsuelo consiste en una capa superior no consolidada, una capa intermedia de Granito fracturado y roca base metamórfica, y un protolito inferior. Dado que tanto las mediciones Como las interpretaciones Tienen incertidumbre, se llevó a cabo un análisis de sensibilidad Para evaluar la incertidumbre del Modelo conceptual, que sugiere lo siguiente (1) Para predecir la descarga del arroyo en la Salida, los parámetros más influyentes son el coeficiente de Manning, el DEM, la hidroestratigrafía y la conductividad hidráulica, y la cubierta terrestre. En comparación con un DEM de menor resolución, un DEM Lidar puede conducir a predicciones más precisas de la descarga del arroyo y del perfil de elevación del arroyo. (2) Para predecir la humedad del suelo, los parámetros más influyentes son la hidroestratigrafía y las conductividades hidráulicas y porosidades asociadas. (3) sobre la base de un ejercicio de calibración, los valores probables de conductividad hidráulica subsuperficial son ~10^–5 m/s (la capa no consolidada), ~10^–6 m/s ( roca madre fracturada), y ~10^–6 m/s o inferiores (protolito).

摘要

山源区流域的综合建模通常受到水文特征研究和监测数据不足的限制。对于美国怀俄明州Medicine Bow 山脉无名流域，本研究整合了区域表层和地下水文与地球物理测量数据，创建了三维综合水文模型，以阐明地表水、土壤水和地下水之间的相互作用。用于构建和校准综合模型的数据包括:数字高程模型(DEM)、流域出口处的河流流量、土壤湿度数据、气象数据以及包括地震折射、空载电阻率和核磁共振(NMR)在内的地球物理调查数据。基于地球物理测量的解释，地下水层理学划分包括一个顶部非固结层、一个由破碎花岗岩和变质岩床组成的中间层和一个下部原岩。鉴于测量和解释都存在不确定性，进行了敏感性分析以评估概念模型的不确定性，结果提示如下:(1)对于预测无名排泄口的河流流量，最有影响力的参数是曼宁系数、DEM、地层层序和渗透系数以及土地覆盖。与较低分辨率的DEM相比，激光雷达DEM可以更准确地预测河流流量和河流高程剖面。(2)对于预测土壤湿度，最有影响力的参数是地层层序和相关的渗透系数和孔隙度。(3)根据校准结果，无名流域的地下渗透系数的可能值约为10^–5 m/s(非固结层)、10^–6 m/s(破碎岩床)和10^–6 m/s或更低(原岩)。

Resumo

A modelagem integrada de bacias hidrográficas de cabeceiras em ambientes montanhosos é frequentemente limitada pela falta de caracterização hidrológica e dados de monitoramento. Para a bacia hidrográfica No-Name, nas montanhas Medicine Bow em Wyoming (EUA), essa pesquisa integra medidas hidrológicas e geofísicas regionais de superfície e subsuperfície para criar modelos hidrológicos tridimensionais integrados onde as interações entre a água superficial, água do solo e água subterrânea são elucidadas. Os dados utilizados para construir e calibrar o modelo integrado incluem: um modelo digital de elevação (MDE), descarga de fluxo medida na saída da bacia, umidade do solo, dados meteorológicos e levantamentos geofísicos, incluindo método sísmico de refração, resistividade aérea e ressonância magnética nuclear (RMN). Com base na interpretação das medidas geofísica, a hidroestratigrafia de subsuperfície consiste de uma camada superior não consolidada, uma camada intermediária de granito fraturado e rochas metamórficas e um protólito inferior. Uma vez que tanto as medições quanto as interpretações possuem incertezas, uma analise de sensibilidade foi realizada para avaliar as incertezas do modelo conceitual, o que sugere o seguinte: (1) Para prever a descarga de fluxo na saída da bacia, os parâmetros mais influentes são o coeficiente de Manning, MDE, hidroestratigrafia e condutividade hidráulica, e a cobertura da terra. Comparado com um MDE de menor resolução, o MDE feito por Lidar pode levar a precisões mais precisas da descarga e do perfil de elevação do rio. (2) Para prever a umidade do solo, os parâmetros mais influentes são a hidroestratigrafia, além das condutividades hidráulicas e porosidade associadas. (3) Com base em um exercício de calibração, os valores prováveis para a condutividade hidráulica de subsuperfície na bacia hidrográfica No-Name são ~10^–5 m/s (camada inconsolidada), ~10^–6 m/s (rocha fraturada), e ~10^–6 m/s ou inferior (protólito).

Appendix: geophysical data

NMR data: A surface NMR dataset was acquired at site W with a GMR (Vista Clara, Mukilteo, WA) using a 45-m circle figure-eight loop with a 20-ms excitation pulse and loop declination 80° east of north. Sixteen stacks were acquired and the data was preprocessed with an 80-Hz bandpass and 500-ms total record length. A multiexponential inversion was performed using GMR1DINV (Vista Clara, Mukilteo, WA) to produce water content and transverse relaxation time (T₂*) distribution results (Fig. 6). The maximum depth of investigation (DOI), indicated in Fig. 6, was estimated based on the resolution matrix. The water-table depth at this location was estimated to be 5 m below the ground surface based on the transition from short relaxation times (<100 ms) to longer relaxation times (>100 ms) indicative of the transition from water in unsaturated pores to saturated pores.

Fig. 6

Surface Nuclear Magnetic Resonance (NMR) sounding. a Water content as a function of depth. The depth of investigation (DOI) is indicated by the solid line. b The transverse relaxation time (T₂*) distribution. Interpreted water-table depth (white-dashed line, b) based on transition from short to longer relaxation times

Seismic refraction data. A total of 13 seismic refraction lines were measured and grouped into northwestern and southeastern grids (Fig. 1b). Example datasets from this campaign shown in Fig. 7 were measured using Geode seismometers (Geometrics, San Jose, CA) with 96 geophones at 2.5-m spacing and 10-m hammer shot spacing. Topography was recorded using a level sight and stadia rod. Dataset 1 was measured on 7 July 2014 and dataset 2 was measured on 23 July 2014. The location of 0 m for each line is indicated on Fig. 1 with a black arrow. The datasets were hand-picked and inverted using DW Tomo (Geogiga Technology Corp., Calgary, AB) with a linear gradient starting model based on the travel-time offset curves and a 5:2 horizontal:vertical smoothing ratio. Picking error was estimated at 0.7 ms and up to eight iterations were allowed, resulting in RMSE residuals ~2.5 ms.

Fig. 7

Example seismic refraction tomograms. a Line 1, northwest grid; b Line 2, southeast grid. Isocontours indicate 1,000 and 3,000-m/s p-wave velocities used for interpretation of subsurface structure