Resolution Of Model Parameters Research Articles

Two dimensional joint inversion of direct current resistivity and radiomagnetotelluric data based on unstructured mesh

Using the unstructured mesh, a new two-dimensional joint inversion algorithm has been developed for Radiomagnetotelluric and Direct current resistivity data. The unstructured mesh is generated with triangular cells, whose vertical and lateral lengths increase towards the depths. The Finite Element Method (FEM) has been used in the forward modelling part of the developed joint inversion algorithm. In the previous studies, structured grid-based joint inversion algorithms have been developed using the Finite Difference Method (FDM). In the structured grid-based algorithms, when the mesh is being generated with rectangular cells, the vertical lengths of the cells get bigger towards the depths while the lateral lengths remain constant. With the structured mesh, the undulated surface topography cannot be represented well enough. Also, because of the incompatible aspect ratio of model cell sizes in deeper model sections, the resolution of the model parameters will get smaller and cannot be resolved well with the structured grids. Imaging of surface topography and underground resistivity structures by the new algorithm requires fewer elements than those using structured grids. Therefore, the developed algorithm is faster than traditional 2D inversion algorithms. Furthermore, the resolution of the deeper model parameters has been increased by using the definition of the unstructured grid. A regularized inversion scheme with a smoothness-constrained stabilizer has been employed to invert the data. First, we have tested the developed joint inversion algorithm using synthetic data simplified from archaeological and mine site scenario and the results have been compared with the conventional algorithms using structured grids. We have also tested our algorithm with the real data which were collected from mineral investigation site at approximately 10 km east of the Elbistan district of Kahramanmaraş province, in the west of the Taurus Mountains, Turkey. The results show that the developed joint inversion algorithm is a powerful tool to detect both resistive and conductive targets.

Re‐parameterisations of the Cole–Cole model for improved spectral inversion of induced polarization data

ABSTRACTThe induced polarization phenomenon, both in time domain and frequency domain, is often parameterised using the empirical Cole–Cole model. To improve the resolution of model parameters and to decrease the parameter correlations in the inversion process of induced polarization data, we suggest here three re‐parameterisations of the Cole–Cole model, namely the maximum phase angle Cole–Cole model, the maximum imaginary conductivity Cole–Cole model, and the minimum imaginary resistivity Cole–Cole model. The maximum phase angle Cole–Cole model uses the maximum phase φmax and the inverse of the phase peak frequency, τφ, instead of the intrinsic charge‐ability m0 and the time constant adopted in the classic Cole–Cole model. The maximum imaginary conductivity Cole–Cole model uses the maximum imaginary conductivity instead of m0 and the time constant τσ of the Cole–Cole model in its conductivity form. The minimum imaginary resistivity Cole–Cole model uses the minimum imaginary resistivity instead of m0 and the time constant τρ of the Cole–Cole model in its resistivity form.The effects of the three re‐parameterisations have been tested on synthetic time‐domain and frequency‐domain data using a Markov chain Monte Carlo inversion method, which allows for easy quantification of parameter uncertainty, and on field data using 2D gradient‐based inversion. In comparison with the classic Cole–Cole model, it was found that for all the three re‐parameterisations, the model parameters are less correlated with each other and, consequently, better resolved for both time‐domain and frequency‐domain data. The increase in model resolution is particularly significant for models that are poorly resolved using the classic Cole–Cole parameterisation, for instance, for low values of the frequency exponent or with low signal‐to‐noise ratio. In general, this leads to a significantly deeper depth of investigation for the , , and parameters, when compared with the classic m0 parameter, which is shown with a field example. We believe that the use of re‐parameterisations for inverting field data will contribute to narrow the gap between induced polarization theory, laboratory findings, and field applications.

Two-dimensional joint inversion of Magnetotelluric and local earthquake data: Discussion on the contribution to the solution of deep subsurface structures

Joint inversion of data sets collected by using several geophysical exploration methods has gained importance and associated algorithms have been developed. To explore the deep subsurface structures, Magnetotelluric and local earthquake tomography algorithms are generally used individually. Due to the usage of natural resources in both methods, it is not possible to increase data quality and resolution of model parameters. For this reason, the solution of the deep structures with the individual usage of the methods cannot be fully attained. In this paper, we firstly focused on the effects of both Magnetotelluric and local earthquake data sets on the solution of deep structures and discussed the results on the basis of the resolving power of the methods. The presence of deep-focus seismic sources increase the resolution of deep structures. Moreover, conductivity distribution of relatively shallow structures can be solved with high resolution by using MT algorithm. Therefore, we developed a new joint inversion algorithm based on the cross gradient function in order to jointly invert Magnetotelluric and local earthquake data sets. In the study, we added a new regularization parameter into the second term of the parameter correction vector of Gallardo and Meju (2003). The new regularization parameter is enhancing the stability of the algorithm and controls the contribution of the cross gradient term in the solution. The results show that even in cases where resistivity and velocity boundaries are different, both methods influence each other positively. In addition, the region of common structural boundaries of the models are clearly mapped compared with original models. Furthermore, deep structures are identified satisfactorily even with using the minimum number of seismic sources. In this paper, in order to understand the future studies, we discussed joint inversion of Magnetotelluric and local earthquake data sets only in two-dimensional space. In the light of these results and by means of the acceleration on the three-dimensional modelling and inversion algorithms, it is thought that it may be easier to identify underground structures with high resolution.

A new perspective of survival data on clinical epidemiology: introduction of competitive risk model

Competing risks occur frequently in the analysis of survival data that should be dealt with competing risk models. Competing risk is an event whose occurrence precludes the occurrence of the primary event of interest. Previous commonly used Kaplan-Meier method tends to overestimate the cumulative survival functions, while the traditional Cox proportional hazards model falsely evaluates the effects of covariates on the hazard related to the occurrence of the event. There are few domestic reports mentioning the concept, application and methodology of competing risk model as well as the implementation procedures or resolution of model conditions and parameters. The current work aims to explain the core concept and methodology of the competing risk model and to illustrate the process of analysis on cumulative incidence rate, using both the cause-specific hazard function model and the sub-distribution hazard function model. Software macro code in SAS 9.4 is also provided to assist clinical researchers to further understand the application of the model so to properly analyze the survival data.

Evidence of serpentinized mantle beneath a non-transform discontinuity at an ultra-slow spreading ridge from wide-angle ocean bottom seismometer data

Non-transform discontinuities (NTDs) are frequently found on ultra-slow spreading ridges. They have crustal structures that differ significantly from those of neighboring neo-volcanic ridges (NVRs). For an ultra-slow spreading ridge with weakened lithosphere and thinned crust, NTDs play an important role in understanding the mechanisms of ridge spreading, hydrothermal venting and mineral resource deposition. In this paper, we present an extensive ocean bottom seismometer (OBS) experiment that was carried out in 2010 at the Southwest Indian Ridge (SWIR) during R/V Dayang Yihao cruise DY115-21. The seismic source was a four-air-gun array with a total volume of or 98.32 L (6000 cubic inches) and a seismic trace spacing of ~250 m. A total of 10832 shots were fired and 52 survey lines (about 2650 km in length) were acquired. Thirty-eight OBSs, each containing a three-component geophone and a hydrophone, successfully recovered information from the crust and upper mantle in NVR and NTD settings. In this paper, we use two wide-angle seismic profiles and a ray-tracing method to constrain the velocity structure of the crust and upper mantle of the NTD between segments 28 and 29 of the SWIR. The processing methods include relocations of shot and OBS positions, time drift corrections and filtering. The widely used ray-theoretical travel-time inversion codes of Zelt and Smith (1992) were used for the modeling. The method includes a damped least square inversion of picked travel times, and we computed the statistical misfit, resolution and uncertainty of model parameters to ensure that the final velocity model had the best fit to our data. The main conclusions are as follows: (1) NTD has a thinner crust (3.2-4.5 km), with an oceanic layer 2 that is 2 km thick and an oceanic layer 3 that is very thin (<1 km) or absent. (2) Oceanic layer 2 has significant lateral velocity variation, indicating the presence of various faults. (3) The uppermost mantle has low velocities (7.2 km/s), indicative of large and deep faults or the presence of serpentinization. We suggest that the low velocity zone in the upmost mantle, together with the thinner crust (<5 km) of the NTD and the presence of faults, result from large-scale serpentinization of the uppermost mantle, providing a good example for investigations of the difference between the Moho discontinuity and a serpentinization front.

Joint Inversion of RMT and TEM Data for Mapping Active Faults in Northern Greece

Northern Greece is an area with one of the most seismically active region in Europe. Several earthquakes occurred during 20th century. The largest earthquake in recent time with magnitude of 6.5 happened in June 1978. The objective of the presented study is to investigate local geological structure and the top of the basement of the Volvi Basin, Northern Greece. The geophysical surveys are carried out using electromagnetic methods, Radiomagnetotelluric (RMT) and Transient electromagnetic (TEM). The investigations are limited to near surface studies with shallow depth (< 200 m). It is expected that the complex geological structure and the vertical distribution of resistivity layers can be defined by RMT and TEM methods. Joint inversion can increase the number of important model parameters and it can also decrease the ambiguity of the model. The result of joint inversion was in agreement with assumption as the RMT data are resolved the surface layer at depth down to 40 m, and TEM data is small enough to resolve the surface structure (z < 10 m), however it can resolve the deeper structure down to 200 m of depth, depending on the resistivity structure. The joint and sequential models can both represent 1-D single inversion of RMT and TEM data. The model parameters are well resolved in joint inversion in comparison with the single RMT and TEM inversions. Hence, the joint inversion of RMT and TEM data will produce good resolution of model parameters in shallow and deeper parts of fault structure.

Spatial resolution considerations for urban hydrological modelling

Summary Hydrological model simulations can be applied to evaluate the performance of low impact development (LID) tools in urban areas. However, the assessment for large-scale urban areas remains a challenge due to the required high spatial resolution and limited availability of field measurements for model calibration. This study proposes a methodology to parameterize a hydrological model (SWMM) with sufficiently high spatial resolution and direct accessibility of model parameters for LID performance simulation applicable to a large-scale ungauged urban area. Based on calibrated high-resolution models for three small-scale study catchments (6–12 ha), we evaluated how constraints implied by large-scale urban modelling, such as data limitations, affect the model results. The high-resolution surface representation, resulting in subcatchments of uniform surface types, reduced the number of calibration parameters. Calibration conducted independently for all catchments yielded similar parameter values for same surface types in each study catchment. These results suggest the applicability of the parameter values calibrated for high resolution models to be regionalized to larger, ungauged urban areas. The accessibility of surface specific model parameters for LID simulation is then also retained. Conducted perturbations in spatial resolution through sewer network truncation showed that while the runoff volume was mostly unaffected by resolution perturbations, lower resolutions resulted in over-simulation of peak flows due to excessively rapid catchment response to storm events. Our results suggest that a hydrological model where parameter values are adopted from high-resolution models and that is developed based on a minimum conduit diameter of 300 mm provides good simulation performance and is applicable to large-scale urban areas with reasonable effort.

3차원 전기비저항 모니터링 자료의 시간경과 역산

모델변수의 증감 및 분해능을 고려하는 새로운 교차모델 제한자를 사용한 시간경과 역산 알고리듬을 개발하였다. 증감을 고려한 교차모델 제한자는 같은 변화비를 갖는 모델변수에 같은 제한을 가한다. 이 교차모델 제한자는 모델변수의 증감에 관계없이 의미 있는 변화를 보이는 모델변수들을 강조할 수 있다. 분해능 교차모델 제한자는 분해능이 낮은 모델변수에 작은 제한을 가하는 반면, 분해능이 높은 모델변수에는 큰 제한을 가한다. 따라서 분해능이 낮은 모델변수도 시간 경과에 따라 의미 있는 변화를 보일 경우 효과적으로 식별할 수 있다. 3차원 전기비저항 모니터링 자료의 수치실험을 통하여 이들 두 교차모델 제한자의 타당성을 검증하였다. 최종적으로, 방조제의 안정성 평가를 위하여 서해안에 위치한 방조제에서 얻은 3 차원 전기비저항 모니터링 자료의 처리에 개발된 시간경과 역산법을 적용하였다. 시간경과 역산의 결과 모니터링이 수행된 기간 동안 방조제에 주목할 만한 변화가 없는 것으로 해석되었다. We developed a time-lapse inversion using new cross-model constraints based on change ratio and resolution of model parameters. The cross-model constraint based on change ratio imposes the same penalty on the model parameters with equal change ratio. This constraint can emphasize the model parameters with significant change regardless of their increase or decrease. The resolution cross-model constraint imposes a small penalty on the model parameters with poor resolution, but a large penalty on the model parameters with good resolution. Thus, the model parameter with poor resolution can be effectively identified in the inversion result if they are significantly changed with time. Through the numerical tests for 3D resistivity monitoring data sets, the performance of these two cross-model constraints was confirmed. Finally, for the safety estimation of a sea dyke, we applied the developed time-lapse inversion to the 3D resistivity monitoring data that were acquired at a sea dike located in western coastal area of Korea. The result of time-lapse inversion suggested that there were no significant changes at the sea dike during the monitoring period.

Characterization of an aquifer in Roorkee, India using the spatially constrained inversion of in‐loop TEM data

The study area is a sewage farm in Saliyar in the state of Uttarakhand, India, which is irrigated using untreated sewage from the nearby city of Roorkee. Previous studies showed that due to sewage irrigation the first unconfined aquifer at a depth of 5–15 m is contaminated. The aim of this study is to characterize the deeper aquifers to a depth of 100 m, underlying the first unconfined aquifer. Therefore 46 in‐loop TEM measurements were carried out in the farming area and an uncontami‐nated reference site with a transmitter loop of and receiver loops of and . To benefit from the geologically expected one‐dimensional subsurface we applied the spatially constrained inversion technique on the data resulting in a geologically reasonable model and allowing the identification of a conducting layer of 28 Ωm as the second aquifer. The resolution of model parameters was computed using singular value decomposition. The low resistivity of the underlying aquifer, however, is not locally restricted to the sewage farm but is also visible in the reference area; it is therefore unlikely caused by irrigation with untreated sewage.

Resolving model parameter values from carbon and nitrogen stock measurements in a wide range of tropical mature forests using nonlinear inversion and regression trees

Objectively assessing the performance of a model and deriving model parameter values from observations are critical and challenging in landscape to regional modeling. In this paper, we applied a nonlinear inversion technique to calibrate the ecosystem model CENTURY against carbon (C) and nitrogen (N) stock measurements collected from 39 mature tropical forest sites in seven life zones in Costa Rica. Net primary productivity from the Moderate-Resolution Imaging Spectroradiometer (MODIS), C and N stocks in aboveground live biomass, litter, coarse woody debris (CWD), and in soils were used to calibrate the model. To investigate the resolution of available observations on the number of adjustable parameters, inversion was performed using nine setups of adjustable parameters. Statistics including observation sensitivity, parameter correlation coefficient, parameter sensitivity, and parameter confidence limits were used to evaluate the information content of observations, resolution of model parameters, and overall model performance. Results indicated that soil organic carbon content, soil nitrogen content, and total aboveground biomass carbon had the highest information contents, while measurements of carbon in litter and nitrogen in CWD contributed little to the parameter estimation processes. The available information could resolve the values of 2–4 parameters. Adjusting just one parameter resulted in under-fitting and unacceptable model performance, while adjusting five parameters simultaneously led to over-fitting. Results further indicated that the MODIS NPP values were compressed as compared with the spatial variability of net primary production (NPP) values inferred from inverse modeling. Using inverse modeling to infer NPP and other sensitive model parameters from C and N stock observations provides an opportunity to utilize data collected by national to regional forest inventory systems to reduce the uncertainties in the carbon cycle and generate valuable databases to validate and improve MODIS NPP algorithms.

Improvement in 1D TDEM Interpretation by Simultaneous Inversion of Data from Two Loop Configurations

This paper presents simultaneous inversion performed on data from central loop and offset receiver TDEM soundings. It is demonstrated that central loop and offset soundings complement each other and that the simultaneous 1D inversion of both data sets (1) yields rapid convergence and better resolution of the model parameters, (2) reduces the importance of distortion caused by induced polarization on interpretation, and (3) provides a good indication of the subsurface geometry in the measurement zone. Simultaneous inversion of central loop and different offset data sets are initially performed over a layered earth. The simultaneous inversion drastically improves interpretation obtained from the long offset data alone, but yields no noticeable improvement for the short offset sounding interpretation. Individual and simultaneous interpretation of central loop and offset soundings are then compared over a layered earth with a polarizable surface layer. The simultaneous interpretation reduces the importance of IP distortion affecting the central loop data and yields a more reliable model. Finally, central loop and offset data are calculated over a 3D conductive body embedded in layered host and individual and simultaneous inversions are performed. Lack of convergence and mediocre fit in simultaneous inversion indicate the presence of lateral conductivity variations. The simultaneous inversion is applied to two field examples from data collected in the Chalco Sub-Basin (Mexico). The first example consists in simultaneous inversion of two neighboring central loop soundings over an earth with no significant lateral conductivity changes. A slight enhancement in the resolution of model parameters is observed, but no substantial improvement in terms of misfit error or number of iterations is achieved compared to the single-site inversion. The second example shows simultaneous inversion of central loop and offset TDEM soundings when lateral conductivity changes occur. The combination of the two soundings never yields a final model that satisfies both data sets, showing the presence of a 2D or 3D heterogeneity in the vicinity of the two sounding sites.

3D FD resistivity modelling and sensitivity analyses applied to a highly resistive phonolitic body

A highly resistive phonolitic body near Sainerholz / Westerwald in Germany has been investigated for geological mapping using vertical electrical soundings in a Schlumberger configuration. Because of its explicitly three‐dimensional shape, conventional 1D and 2D interpretation techniques are not applicable. Therefore, a new 3D finite‐difference forward modelling algorithm has been applied to acquire information about its subsurface structure and to explain the observed data. This investigation focuses on two exposed soundings: one located near the centre of the body and the other close to its rim. For the interpretation, data from electromagnetic measurements on the lateral extension of the body are additionally taken into account as well as geological a priori information. A possible 3D conductivity model is presented and evidence for its validity is discussed using model studies and sensitivity analyses. The latter are carried out using a newly developed 3D FD sensitivity modelling code with which the total subsurface response can be decomposed. This permits the determination of the resolution of model parameters, indicating the contribution of different parts of the model to the overall response. The results emphasize the feasibility of 3D forward modelling in practice.

Lateral variations of conductivity structure across Southern Scotland and Northern England

High-frequency (100-0.01 Hz) magnetotelluric (MT) soundings have been carried out along a 140 km profile across Southern Scotland and Northern England, where previously longer-period soundings had been undertaken. Examples of the results of the processing of data from 15 sites along the profile and of the application of a number of 1D inversion procedures are presented. The latter include the joint inversion of d.c. resistivity and MT data at Rookhope where borehole stratigraphy is also available. Although the data were severely contaminated by near field sources at some sites, and the station spacing was too large in parts of the traverse for good resolution of model parameters, by incorporation of results from the earlier studies it is possible to obtain a preliminary 2D model of the electrical resistivity structure of the upper crust in this region. Significant differences have been found in the three main geological zones crossed by the survey—the Alston Block, the Northumberland Basin and the Southern Uplands of Scotland. In the discussion, special attention has been paid to four sites in the neighbourhood of the Southern Uplands Fault, for which the data have been subjected to tensor decomposition. This has revealed two dominant regional azimuths, one of which dominates at the shorter periods and corresponds to the strike of the Southern Uplands Fault and the other, at longer periods, which corresponds to the strike of the well-known Southern Uplands magnetovariational (MV) anomaly. Moreover, in this region, the electrical resistivity models indicate the presence of a highly conducting zone beginning at a depth of about 4 km at Station CAP, which is located on the axis of the MV anomaly, and extending to a depth of at least 25 km. The depth to this conducting zone increases both to the north and the south of the MV anomaly axis, resulting in a structure which has many features in common with those along parallel traverses in SE Scotland and in Ireland. In discussing the implications of the Irish model, another researcher proposed either metamorphosed sediments and/or serpentinised island are crust as being possible sources of the high conductivity and other geophysical and geological data. The probability that the arguments in that study are equally applicable to Southern Scotland is currently under consideration.

Analysis of sparse wide-angle reflection data in the τ(p) domain

Uncertainties in ray parameter p for events recorded with small-aperture arrays can be quite large and lead to sizable second-order errors in intercept time τ when source gathers are slant stacked into the τ- p domain. The problem becomes significant for data sets characterized by large gaps in offset and ray parameter coverage. To estimate the uncertainty in ray parameter p arising from small recording apertures, we use a statistical test to evaluate the cross-correlation of a predicted array response with the observed response extracted from the slant stack. The cross-correlation is carried out using the projections of the array responses along the ray parameter axis, following the linear trajectory defined by the average offset over which the slant stack was computed. The method adjusts automatically to the bandwidth of the arrival and flags arrivals for which the ray parameter is poorly resolved because of either high noise levels or overlapping with other arrivals in the stack. We then use the uncertainties in ray parameter to estimate second-order errors in τ. The latter are inversely proportional to curvature in the travel time ( T ) - offset ( X ) domain, or equivalently, are directly proportional to curvature in the τ- p domain. For reflections, therefore, the errors increase with angle of incidence. Errors for S -wave arrivals are smaller than errors for P -wave arrivals for a given ray path because of the stronger curvature of the S -wave travel-time branches. Second-order errors in τ for quarry blast data recorded with small-aperture arrays in eastern Pennsylvania range from 0.05 to 0.4 sec, and in general are greater than errors due to uncertainties in origin times and finite bandwidth. The uncertainties in τ associated with limited-aperture recordings therefore can have a significant effect on the tradeoff between variance and resolution of model parameters when the data are inverted for velocity structure.

Tomographic inversion of local earthquake data from the Hengill‐Grensdalur Central Volcano Complex, Iceland

We have determined the three‐dimensional P wave velocity structure within the area of the Hengill‐Grensdalur central volcano complex, southwest Iceland, from the tomographic inversion of 2409 P wave arrival times recorded by a local earthquake experiment. The aperture of the 20‐element seismic network utilized in the inversion permitted imaging of a 5‐km‐thick crustal volume underlying a 15×14 km2 area. Within this localized volume are located the underpinnings of the active Hengill volcano and fissure swarm, the extinct Grensdalur volcano, and an active high‐temperature geothermal field. It was thus expected that the characteristic length scale of heterogeneity would be of the order of a kilometer. In order to image heterogeneous seismic velocity structure at this scale we paid particular attention to the fidelity of the assumed model parameterization, defined as the degree to which the parameterization can reproduce expected structural heterogeneity. We also discuss the trade‐off between the resolution of model parameters and image fidelity, compare results obtained from different parameterizations to illustrate this trade‐off, and present a synoptic means of assessing image resolution that utilizes the off‐diagonal information contained within the resolution matrix. The final tomographic image presented here was determined for a parameterization with fidelity that closely matches the geologic heterogeneity observed on the surface. For this parameterization, the resolution of individual parameters is generally low; however, a quantitative analysis of resolution provides an unambiguous assessment of well‐resolved volumes. Within the better resolved regions of the model the averaging volumes are 1–2 km and 2–4 km in vertical and horizontal extent, respectively. Results of tomographic inversion image three distinct bodies of anomalously high velocity, two of these extend from near the surface to a depth of about 3 km. These high‐velocity volumes are located directly beneath the surface expressions of the extinct Grensdalur volcano and the extinct Husmuli basalt shield. The third high‐velocity structure occurs in the depth range of 3–4 km but does not extend to the surface. These three high‐velocity bodies are interpreted to be solidified magmatic intrusions. Relatively low velocities underlay limited portions of the trace of the present accretionary axis and a low‐velocity body is imaged in the roots of the active Hengill volcano. The volume of lower velocities located beneath the surface expression of the Hengill volcano is interpreted to be a region of partial melt.

Simultaneous inversion of hydrogeologic and thermal data: 2. Incorporation of thermal data

Introduction of a data set for a second potential field provides an approach to deal with data limitations in the inverse problem. We examine how temperature measurements can be used to improve the resolution of model parameters in constructing a model of a steady state groundwater flow system. The approach is applicable to cross‐sectional models with vertical components of fluid flow. In media with low hydraulic conductivity, the thermal regime is conductive and temperature measurements do not carry direct information on the hydrogeologic regime. In more permeable systems however, fluid velocities are sufficient to redistribute heat within the basin, perturbing the conductive regime. Parameter estimates developed in the inverse simulation are required to satisfy a weighted measure of both the hydrogeologic and thermal data. The effectiveness of this joint inverse depends upon the extent of the advective disturbance, which, in turn, is controlled by the magnitude of hydraulic conductivity, the length scale and location of hydraulic conductivity variations, the depth of active flow, and the regional configuration of the groundwater flow system. A series of hypothetical examples are presented to demonstrate how model misfit can be reduced when noisy hydraulic head data are augmented with thermal data. Because the extent of an advective disturbance depends upon the absolute value of hydraulic conductivity, addition of temperature measurements in an inverse simulation is particularly helpful in cases where there is uncertainty in boundary conditions, such as recharge rates. The approach is used to construct a model of the groundwater flow system at Downie Slide, north of Revelstoke, British Columbia. Results presented in part one suggest that the slide behaves as a homogeneous, isotropic medium (Woodbury et al., 1987). With a homogeneous hydraulic conductivity, only the ratio of fluid flux to hydraulic conductivity could be determined. By including thermal data, the value of hydraulic conductivity can be estimated.