Comparison of Various Growth Curve Models in Characterizing and Predicting Water Table Change after Intensive Mine Dewatering Is Discontinued in an East Central European Karstic Area

Kamilla Modrovits,József Kovács,András Csepregi,Ilona Kovácsné Székely,István Hatvani

doi:10.3390/w13081047

Abstract

The modeling of karst water level fluctuations is a crucial task in the water resource management of vulnerable karstic areas. In the Transdanubian Range (East Central Europe, Hungary), from 1950 to 1990, coal and bauxite mining were carried out, with large amounts of karst water being extracted, thus lowering the water table by amounts ranging between 10 and 100 m. Since the cessation of mining activities in the early 1990s, the volume of natural recharge has exceeded the amount of dewatering, and the system has begun to return to its original undisturbed state. This apparently welcome development does, however, bring economic and technical engineering problems. The estimation and prediction of such water level changes is often tackled via the use of deterministic approaches, however, in the present case, it is also addressed with an alternative approach using trend estimation to monthly water level data from 107 karst water wells over the period 1990–2017. To approximate the change in karst water levels, (i) growth curve models were fitted to the monthly data, allowing the estimation of karst water levels, at least as far as 2030. Similarly, this was also done with (ii) deterministic modelling in order to describe the recovery process up to 2030. Specifically, measured and predicted values for karst water level were used to derive interpolated (kriged) maps to compare the forecasting power of the two approaches. Comparing the results of the trend analysis with those of the traditional deterministic modelling results, it is apparent that the two approaches predict similar spatial distribution of water levels, but slightly different future water level values.

Highlights

Mining has evolved with the development of humankind helping to meet the everincreasing demand for raw materials [1,2]
The grouping was 100% successful according to the linear discriminant analysis and can be well illustrated using the LD1 and LD2 functions (Figures 4 and 5)
The time series of karst water monitoring wells in the Transdanubian Range were split into three groups by cluster analysis, and these groups may be spatially separated on the basis of their location relative to the depression centers, which are characterized by different patterns

Summary

Introduction

Mining has evolved with the development of humankind helping to meet the everincreasing demand for raw materials [1,2]. Large-scale mining activities are often associated with a number of environmental problems, such as air, noise, soil, or water pollution. Other effects of mining, such as the infrastructure required by industrial-scale activity and changes to the landscape, may persist well beyond the cessation of mining operations, even decades after the closure of the mine [3,4]. One of the most common problems is the decrease in water levels associated with over-abstraction, a widespread global phenomenon as a result of increasing extraction in the course of mining operations and for drinking water.

Methods

Results

Discussion

Conclusion