IMPLEMENTING THE SHALE HILLS WATERSHED MODEL IN APPLICATION OF PIHM

ABSTRACT

 

This thesis addresses the problem of model implementation at the Shale Hills experimental watershed using PIHM (Penn State Integrated Hydrologic Model). Ever since the 1970s, interdisciplinary teams have been working in Shale Hills watershed to study a wide range of earth science problems. However, a watershed model was not constructed until Qu and Duffy (2007) proposed the PIHM model. In recent years, the PIHM had a major update by Kumar and Bhatt where they added new flux components to the channel flow, implemented macropore effects, throughfall drainage, evaporation from ground and transpiration from the canopy. At the same time, PIHM was extended to include national databases which are referred to as A-priori data. This research was performed with NSF funding through the Susquehanna River Basin Project, the Critical zone Observatory project and the RTH_NET project. The focus of this thesis is to implement a new version of PIHM at Shale Hills using data sets recently acquired through the Critical Zone Observatory Project. These new data sets include: 3 meter digital elevation data, a new bedrock elevation coverage, the latest soil classification data from SSURGO with site specific extensions to SSURGO made by H. Lin’s group, and the National Land Cover Data distribution dataset. The updated model is calibrated through a trial and error process, using the 1974 artificial irrigation experiment by Lynch et al. The model successfully reproduces the runoff at the watershed outlet during a sequence of 6 rainfall events. It shows that the Horton overland flow and subsurface storm flow are the main drive for the runoff peak in the channel. The model also simulates groundwater levels, recharge, transpiration, etc. The model represents a preliminary calibration which will be implemented in real-time model with current data. Operating the model in realtime will allow the continuous calibration using CZO experimental data and provide feedback to scientists. Finally, Bhatt and Kumar have implemented a GIS interface for PIHM and this was used at Shale Hills for setting up the model new data coverages. A tutorial for the PIHMgis and the PIHM model is included as part of this thesis.

 

 

TABLE OF CONTENTS

 

List of Figures…………………………………………………………………………….viii

List of Tables………………………………………………………………………..……xii

List of Notation…………………………………………………………..…….……..…xiii

Acknowledgements…………….…………………………………………….……..……xvi

 

 

Chapter 1.Introduction…………………………………………………………………………………………..1

1.1 Literature Review…………………………………………………………………1

1.2 Motivation and Objective……………..…………………………………………..3

 

 

Chapter 2. Methodology and Approach…………………………………….….………….5

2.1 PIHM Model………………………………………………………………………5

2.1.1 Semi-discrete Volume Method…………………………………………….5

2.1.2 Domain Decomposition and Fully Coupled Model Architecture……….…7

2.1.3 Physical Processes…………………………………………..………..…….8

2.2 PIHMgis….……………….…………………………..………………………….14

 

 

Chapter 3. A-priori Data Development for Shale Hills Watershed…………………..…..17

3.1 Surface Elevation…………………………………………………………..……18

3.2 Bedrock Elevation……………………………………………………………….19 3.3 Soil Mapping……..……………………………..…………….….…….…………21 Chapter 4. Getting Data Into Model

A Short PIHMgis Tutorial (Shale Hills Watershed)…………………….……25

4.1 Introduction………..……………………………………………………………..25

4.2 Software Initialization…………………………………………………………..26

4.3 Data Preparation..……………………………..…………….….…….…………..27

4.4 Raster Processing…………………………………………………………………27

4.4.1 Fill Pits……………………………………………………………..28

4.4.2 Flow Grid………………………………………………..…………29

4.4.3 Stream Grid…………………………………………………………32

4.4.4 Link Grid……………………………………………………………34

4.4.5 Stream Polyline……………………………..………………………35

4.4.6 Catchment Grid..………………………………………………..….36

4.4.7 Catchment Polygons…………………………………………………38

4.4.8 Identify the Watershed………………………………………………39

4.5 Vector Processing…………………………………………….…………………..40

4.5.1 Polygon To Lines…………………………………………………..41

4.5.2 Correction of the Stream and the Boundary…………………………42

4.5.3 Simplify Line…………………………………..……………………43

4.5.4 Split Line……………………………………………………………45

4.5.5 Vector Merge……………………………………………………….45

4.6 Domain Decomposition..………………………………………………….……..46

4.6.1 Read Shape Topology………………………………………………47

4.6.2 Run TRIANGLE……………………………………………………48

4.6.3 TIN Generation…….………………………………………………50

4.7 Data Model Loader…….…………………………………………………………51

4.7.1 Mesh File Generation………………………………………………52

4.7.2 Attribute File Generation….………………………………………53

4.7.3 River File Generation….……………………………………………55

4.7.4 Parameter File Generation………………………………………….56

4.7.5 Completion of the data model for PIHM……………………………58

4.8 Discussion……….…….………………………………………………….…….58

 

 

Chapter 5. Land Cover Classification using Topographic Wetness Index……………….59

5.1 Introduction ……………………………………………………………………..59

5.2 Wetness Index……………………………………………………………………60

5.3 Wetness Index Calculation and Land Cover Classification.….…….……………60

 

 

Chapter 6. Results…………………………………………………………………………65

6.1 Model Domain and Experiment Setup…………………………………………..65

6.2 Simulation of Stream Flow………………………………………………………66

6.3 Simulation of Soil Moisture……………………………………………………..67

6.4 Simulation of Evaporation and Transpiration……………………………………69

6.5 Water Budgets……………………………………………………………………70

6.6 Calibration Strategy………………………………………………………………70

6.6.1 Calibration Process…………………………………………………………70

6.6.2 Sensitivity to Conductivity…………………………………………………71

Chapter 7 Conclusions and Future Work…………………………………….…………..94

 

References………………………………………………………………………………..96

Chapter 1

Introduction

 

 

1 .1 Literature Review

 

The thesis addressed the problem of constructing a model of Shale Hills experimental watershed using PIHM (Penn State Integrated Hydrologic Model). Mathematical models have been widely accepted as an important tool for theoretical and experimental research in hydrology science in recent years. Duffy (1996) proposed a twostate model by integrating local conservation equation with respect to the soil moisture storage of the saturated and unsaturated states. The model was tested in the Shale Hills watershed, and showed that soil moisture and saturated storage are controlling factors for the hydrologic response of first order watershed (Tchaou, 1999). Qu and Duffy (2007) developed a multi-process, multi-scale, spatially distributed, physically based hydrologic model, in which, major processes (surface overland flow, subsurface flow, channel routing, interception, snowmelt, evaporation and evapotranspiration) are fully coupled using the semi-discrete finite volume method. In the model, the TIN (triangular irregular network) is generated by Delaunay triangulation. The model is solved by an implicit sequential solver from SUNDIALS (suite of nonlinear and differential algebraic equations solver), which is developed in LLNL (Lawrence Livermore National Laboratory). The model is also known as the first version of PIHM (Penn State Integrated Hydrologic Model). Qu and Duffy (2007) implemented the approach to simulate the Shale Hills field experiment, and “successfully simulated observed groundwater levels, as well as runoff at the outlet and internal points within the watershed”. Kumar and Bhatt revised the code, and enhanced it by adding the macropore phenomena, throughfall drainage, evaporation of canopy and ground process to the model, which was the second generation of PIHM code. On the other hand, however, the large number of physical parameters that need to input to the model becomes a big barrier for the wide application for the model (Bhatt et al. 2009). Bhatt (2009) integrated the PIHM model to an open-source Geographical Information System (GIS), which was known as PIHMgis. PIHMgis applies the Qgis interface for the preprocessing of topography, geology, soil, vegetation and climate data, which greatly reduced the effort of using the model.

Shale Hills experimental watershed is one of the Critical Zone Observatories. It is a typical V-shape, forested, small catchment lying in the Valley and Ridge physiographic province of East United States. The Shale Hills experimental watershed is a first-order, 19.8 acre watershed approximately 14 miles southwest of State College in

Barree Township, Huntingdon County, and is part of the Stone Valley Experimental Forest owned by Pennsylvania State University. The climate is transiting through the seasons. Lynch (1976) performed artificial rainfall experiments in the 1974, by the irrigation system installed in the watershed. Water budgets are collected to investigate the effects of antecedent soil moisture on storm flow volumes and timing. Lin (2006) revisited the site in 2003 and 2004 for a better understanding of temporal and spatial distribution of soil moisture at the Shale Hills watershed. Based on local measurements, a

3-m DEM is refined from 10-m DEM for a better representation of the swales and ridges. Lin (2006) also identified the 5 soil classes and detailed the soil map for the Shale Hills experimental watershed. From Lin’s research, Blairton and Ernest are located on the valley floor along the stream, Rushtown is found in the swales, while Berks and Weikert are found on the slopes and ridges. Species of maple, oak, hickory, which are typical deciduous trees in Shale Hills watershed, distribute on the slopes and ridges. Hemlocks are found on the valley floor near the west, Pines are located up on the ridges of the southern slope.

 

 

1.2 Motivation and Objective

 

Since Qu and Duffy (2007) implemented the Shale Hills watershed in the first version of PIHM, the PIHM model had a major update by Kumar and Bhatt. The 1D channel flow is updated by multiple channel flow components; and macropore effects, evaporation from ground and transpiration are newly integrated. It is hoped that more information can be obtained from the model, to help us understand the hydrologic mechanism in Shale Hills area and also helped us in experimental planning.

Although Bhatt had integrated the GIS interface to the PIHM model as its pre-

processing part, the use of the PIHMgis is still quite subtle. As part of the thesis, a tutorial for the PIHMgis and PIHM model is included to improve the wide application of the

PIHMgis and PIHM model to other watersheds.

Newly obtained data from Henry Lin’s group is ready to update the soil map, the surface elevation map and the bedrock elevation map. As the experiments go on in the Shale Hills watershed, the ultimate goal of the work is to establish a real-time hydrologic model to help us understand the hydrologic behavior and facilitate the studies in the watershed.

IMPLEMENTING THE SHALE HILLS WATERSHED MODEL IN APPLICATION OF PIHM

Sharing is caring!