%40تخفیف

Controlling Parameters on Karst Aquifer Hydrodynamic Behavior

تعداد169صفحه در فایل word

Ph.D

Controlling Parameters on

Karst Aquifer Hydrodynamic Behavior

This research is focused on the determination of karst aquifer characteristics by means of spring hydrograph, as well as investigation of controlling aquifer parameters on its hydrodynamic behavior. Using spring hydrograph, novel parsimonious methods were introduced for estimation of mean residence time, MRT, and recharge coefficient, as two useful hydrodynamic parameters for karst aquifer characterization. Effect of different aquifer parameters, on its hydraulic and transport behavior due to a recharge event, was also investigated by means of a state-of-the-art distributed-parameter hybrid flow and transport model, in the framework of global sensitivity analysis. The considered parameters are the characteristics of epikarst diffuse system-unsaturated zone, epikarst vertical quick flow, and saturated zone (conduit and matrix systems, conduit associated drainable storage, as well as exchange coefficient). Sheshpeer karst aquifer, in south-central Iran, was chosen as the case study for the method applications.

MRT of emerged water from a karst spring during recession periods, was estimated providing long-term consecutive spring hydrographs, ranging from high to low flow, and the relevant precipitation hyetographs. The method is based on a discharge-weighted average formula, where the discharge components with known residence times, are estimated by horizontal displacement of segmented-exponential master recession curve, MRC, at the end of measured recessions. The MRT by the proposed method is representative of all active circulating water throughout the entire aquifer, presenting the combined effect of aquifer hydrodynamic characteristics. The proposed method is more appropriate than the costly isotopic methods, for practical applications such as vulnerability assessments, in case a small portion of water with exceptionally high residence time increases the mean.

Horizontal displacement of segmented-exponential MRC at the end of recessions, was further utilized for estimation of actual event recharged water volume, VR_E, and consequently, the event-based recharge coefficient, RC_E, which is an important tool for evaluation and management of karst water resources. It was proved that the calculated VR_E, takes the change in groundwater storage into account, enabling an accurate RC_E estimation.

The enhanced version of the recent hybrid model MODFLOW-2005 Conduit Flow Process Mode 1, namely CFPv2, was considered as the code for modeling studies. Sobolꞌ variance-based global sensitivity analysis, and spearman rank correlation coefficient, were used to investigate three hypothetical idealized aquifer settings with different conduit geometries (i.e. single, branched and networked), but similar distributed and direct recharge components. Results shows that the conduit diameter is by far the main controlling parameter on the spring hydro-chemo-thermo-graphs and potentiometric head in the conduit system, for all the considered models; while matrix head is relied on more than a parameter, governed by the exchange coefficient, saturated / unsaturated matrix conductivities and water contents. Subsequently, simple but realistic models were setup for the Sheshpeer catchment considering different probable conduit geometries based on the available hydro-geological information. Automated calibration of model influential parameters was performed based on the measured spring hydro-thermo-graphs together with a tracer breakthrough curve. Results suggested that successful simultaneous calibration of all the measured spring signatures can be attained just for one of the conduit geometries with highest connectivity to the sinkholes over the catchment. Although the model could reasonably simulate all the spring signatures, and the reported parameter uncertainties were quite low, the calibrated parameter values cannot be accepted unless field measurements on conduit system characteristics (i.e. its configuration and sizes) and matrix hydraulic parameters, as well as long term observation of aquifer head and spring physicochemical signatures be collected.

Keywords: Karst aquifer, recession curve, mean residence time, hybrid models, hydrodynamic and transport behavior

دسته: زبان خارجه, زبان خارجه برچسب: hybrid models, hydrodynamic and transport behavior, Keywords: Karst aquifer, mean residence time, recession curve

توضیحات

Table of Contents

Content Page

1- Objectives and literature review.. 2

1-1- Introduction. 2

1-2- Research objectives and structure of the thesis. 2

1-3- Controlling parameters on the spring hydrograph. 4

1-4- Master recession curves. 8

1-5- Groundwater mean residence time. 10

1-6- Recharge coefficient 12

1-7- Conceptual model of karst aquifers. 16

1-8- Numerical modeling of karst aquifers. 18

1-8-1- MODFLOW CFPv2. 22

1-9- Global sensitivity analysis. 26

1-10- Model calibration. 30

2- Study site: Sheshpeer karst aquifer. 36

2-1- Introduction. 36

2-2- Hydrogeological setting. 37

2-3- Catchment database. 39

2-3-1- Spring hydro-chemo-thermo-graphs. 39

2-3-2- Tracer test 41

2-3-3- Precipitation data. 43

3- Methods for mean residence time and recharge coefficient estimation. 46

3-1- Introduction. 46

3-2- Mean residence time estimation method. 46

3-3- Event-based recharge coefficient estimation method. 49

4- Numerical model setups of karst aquifers. 57

4-1- Introduction. 57

4-2- Model implementation of recharge. 57

4-3- Model representation of idealized karst aquifers. 59

4-3-1- Recharge function. 60

4-3-2- Solute and heat transport setups. 62

4-3-3- Statistical distributions of input parameter 63

4-4- Model representation of the Sheshpeer aquifer 66

4-4-1- Spatiotemporal distribution of recharge. 68

4-4-2- Conduit geometries. 71

4-4-3- Solute transport setup. 71

4-4-4- Heat transport setup. 72

4-4-5- Model calibration procedure. 73

5- Results and discussions. 76

5-1- Introduction. 76

5-2- Estimation of groundwater mean residence time for Sheshpeer aquifer 76

5-3- Estimation of recharge coefficient for the Sheshpeer aquifer 84

5-4- Application of hybrid flow and transport models. 91

5-4-1- Idealized models. 91

5-4-1-1- Base case results. 91

5-4-1-2- Global sensitivity analysis results. 98

5-4-2- Inverse model application for the Sheshpeer aquifer 111

5-4-2-1- Preliminary stage model results. 111

5-4-2-2- Second stage models and results. 113

6- Conclusions. 126

6-1- Introduction. 126

6-2- Groundwater mean residence time. 126

6-3- Event-based recharge coefficient 127

6-4- Hybrid flow and transport models. 128

References. 130

Abstract and Title Page in Persian

List of Figures

Figure Page

Figure ‎1‑1- Distribution of major outcrops of carbonate rocks in (a) The World (Williams and Fong, 2010), and (b) Iran (modified from Raeisi and Kowsar, 1997). 3

Figure ‎1‑2- MRC of the Sheshpeer Spring, which is constructed by matching strip method. The MRC is composed of five exponential segments. Recession coefficient and the R² for the fitted exponential function in each segment is indicated. Note that the vertical axis is logarithmic. 10

Figure ‎1‑3- Schematic representation of a regional hydrogeological system presenting possible flow paths and travel times from the recharge to the discharge areas (Kazemi et al., 2006). 11

Figure ‎1‑4- Schematic representation of a karst aquifer (Mangin, 1994). 17

Figure ‎1‑5- Distributed parameter modeling approaches for karst aquifers (Kovács and Sauter, 2007). 19

Figure ‎1‑6- Conceptual and numerical representation of hybrid models (Birk et al., 2005a). 21

Figure ‎1‑7- Conceptual and model representations of CADS in MODFLOW CFPv2 (modified from Reimann et al., 2013). 25

Figure ‎2‑1- Sheshpeer Spring outlet. 36

Figure ‎2‑2- Hydrogeological map of the study area. 38

Figure ‎2‑3- Measured hydro-chemo-thermo-graphs of the Sheshpeer Spring. 40

Figure ‎2‑4- Measured and reconstructed discharge of the Sheshpeer Spring along with recorded precipitation at the Berghan meteorological station. 41

Figure ‎2‑5- Tracer breakthrough curve for the dye trace during the 1992 recession period, plotted along with the spring discharge. 42

Figure ‎2‑6- Spatial distribution of precipitation gauge stations on the TIN map. Sheshpeer Spring and its catchment are indicated. 44

Figure ‎3‑1- Diagrams showing the principle of the MRT estimation method, using hypothetical discharge curves and precipitation events. 48

Figure ‎3‑2- Diagrams showing the principle of the event-based RC estimation method, using schematic discharge curves and precipitation events. 51

Figure ‎4‑1- Schematic cross sectional view of the model recharge representation for karst aquifer. 59

Figure ‎4‑2- Model representation of idealized aquifers with conduit geometries A, B, and C. 61

Figure ‎4‑3- Probability plots for the lognormal distributions fitted to the worldwide tortuosity data from Worthington (2015). 65

Figure ‎4‑4- (a) Model grid of the Sheshpeer aquifer; (b) Elevation contour defining the top of model; (c) Distribution of concentrated and distributed recharge cells in the model domain, which were defined by RCH1 and UZF1 MODFLOW-2005 Packages, respectively. 67

Figure ‎4‑5- Time series of estimated mean daily values of maximum air temperature, precipitation, and recharge, over the Sheshpeer catchment. 70

Figure ‎4‑6- Selected conduit geometries for the Sheshpeer aquifer: (a) I, (b) II, (c) III, and (d) IV. 72

Figure ‎5‑1- Measured discharge at Sheshpeer Spring for 1990–1992 (solid line) and discharge inferred from other data for 1979–1989 (dotted line). 77

Figure ‎5‑2- MRC of Sheshpeer Spring on a semi-log scale. 77

Figure ‎5‑3 Measured (solid lines) and extrapolated (dash lines) recession curves of the Sheshpeer Spring. Some parameters of MRT calculation for individual hydrograph “a” were indicated on June 22, 1992 (cf. Figure 5-1 and 5-2). 80

Figure ‎5‑4- Residence-time distribution on June 22, 1992, based on the decomposing of total discharge on June 22, 1992, into discharge components by the method described in the text and shown in Figure 5-3. 81

Figure ‎5‑5- The colored curves show the MRT computed for the outflow from Sheshpeer Spring at different times during the second recession of 1992 for different assumptions about the mean age in years of the earliest discharge component. 83

Figure ‎5‑6- Sheshpeer Spring hydrograph and the synchronous precipitation hyetograph at the Berghan station. 85

Figure ‎5‑7- Spatial variation of mean annual precipitation over the Sheshpeer catchment during 1972 to 2012. 87

Figure ‎5‑8- Estimated RC_E (gray bars) and RC_Y (Black bars) for the Sheshpeer aquifer during hydrological years 1990-1991 to 1992-1993. 89

Figure ‎5‑9- Simulated spring (a) hydrograph, (b) chemograph and (c) thermograph for the idealized models A, B, and C. 92

Figure ‎5‑10- Time series of (a) conduit head, h_c; (b) matrix head at the conduit node, h_m₁; and (c) matrix head at the matrix cell, h_m₂, for the models A, B, and C. 95

Figure ‎5‑11- CADS and matrix discharge versus the matrix and conduit head at the sinkhole node of the model A. 96

Figure ‎5‑12- Normalized chemo-thermo-graphs for the idealized models. 98

Figure ‎5‑13- Scatterplots of chemograph standard deviation (chemograph SD) versus conduit diameter (d_c), CADS width (W_CADS), and conduit direct recharge (CRCH), for the model A. 102

Figure ‎5‑14- Total order index of model parameters in relation to different statistical indicators of (a) spring and (b) aquifer head signatures, for idealized model A. 103

Figure ‎5‑15- Interaction index of model parameters in relation to different statistical indicators of (a) spring and (b) aquifer head signatures, for idealized model A. 104

Figure ‎5‑16- Total order index of parameters with respect to different statistical indicators of aquifer signatures, for the idealized model A. 105

Figure ‎5‑17- Absolute values of SRCC for the parameters in relation to different spring signature indicators for the idealized models (a) A, (b) B, and (c) C. 108

Figure ‎5‑18- Absolute SRCC values for the parameters in relation to different aquifer head indicators for the idealized models (a) A, (b) B, and (c) C. 109

Figure ‎5‑19- Measured (black dotted lines) and simulated (continuous colored lines) Sheshpeer Spring signatures for the models I (red), II (orange), III (green), and IV (blue). 112

Figure ‎5‑20- Structuring of conduit diameter for the model IV: (a) IV-a, (b) IV-b (c) IV-c, (d) IV-d, and (e) IV-e. 115

Figure ‎5‑21- Model selection criteria (AIC, AICc, BIC, KIC) and weighted least squares objective function (Φ) values for different calibrated models. 116

Figure ‎5‑22- 95% confidence intervals (black thin bars), and limits of reasonable ranges of parameter values (grey thick bars), plotted as the percentage of estimated values for the model IV-c. 119

Figure ‎5‑23- Measured (black dotted lines) and simulated (continuous blue lines) Sheshpeer Spring signatures, for the model IV-c. 121

Figure ‎5‑24- (a) Histogram and (b) Cumulative distribution function of weighted residuals versus the theoretical normal distribution function. 122

Figure ‎5‑25- Plot of weighted measured versus simulated observations, containing spring discharge “Q”, solute concentration “C”, and water temperature “T”, for the model IV-c. 122

List of Tables

Table Page

Table ‎1‑1- Controlling parameters on karst aquifer recession curves according to the literature. 5

Table ‎4‑1- Incorporated parameters in the sensitivity analysis. 64

Table ‎5‑1- Recession coefficients, discharge ranges, and durations of the exponential segments for the measured recession curves for the years 1990–1992 compared to the parameters of the segmented MRC. 79

Table ‎5‑2- The approximate calculated MRTs for June 22, 1992, and throughout the duration of the recession, for various assumed mean ages for the oldest discharge component. 82

Table ‎5‑3- Achieved MLR models for the events, hydrological years and long-term annual mean of 1972 to 2012. 86

Table ‎5‑4- Estimated RC_E and RC_Y for Sheshpeer Spring during hydrological years 1990-1991 to 1992-1993. 88

Table ‎5‑5- Rank of parameters with total order index of over 0.1 for the model A. 106

Table ‎5‑6- Rank of parameters with |SRCC| of over 0.2 for the model A. 110

Table ‎5‑7- Rank of parameters with |SRCC| of over 0.2 for the model B. 110

Table ‎5‑8- Rank of parameters with |SRCC| of over 0.2 for the model C. 110

Table ‎5‑9- RMSE of different model observations for all models. 112

Table ‎5‑10- Estimated parameter values and their 95% confidence intervals for models with different conduit structures. 118

Table ‎5‑11- Some measures of overall fit goodness for the model IV-c, reported by PEST. 120

Table ‎5‑12- Some measures of goodness of fit for individual model observations of model IV-c. 123

امتیاز 0 از 5 از 0 دیدگاه

قیمت اصلی 35,000 تومان بود.قیمت فعلی 21,000 تومان است.

تاریخ ایجاد دی 5, 1399
تاریخ بروزرسانی اسفند 13, 1401
فروش 0
دیدگاه 0

شرایط استفاده از محصول

تمام محصولات به ازای پرداخت پروژه تنها قابل استفاده می باشند. این بدین معنی است که برای استفاده از یک یا چند محصول ، لازم به خرید آن محصول است.

مزایای خرید این محصول:

دسترسی به فایل به صورت همیشگی
پشتیبانی رایگان
پشتیبانی 24 ساعته محصول
بازگشت وجه در صورت عدم رضایت مشتری

مشاهده قوانین سایت