Keywords

1 Introduction

More than 95% of floodplains and half of the wetlands have been converted to urban and agricultural land [1]. Consequently, the availability of water in floodplain has been changed. The incision of the Drava River leads to groundwater levels 1.5 to 2.5 m lower and deteriorated the water budget of the floodplain [2, 3]. Moreover, human interventions (river regulation, hydroelectric plants, etc.) generated drought hazard and water shortage in the floodplains of Hungary [3], jeopardizing agricultural production and ecosystem services [4]. Floodplain rehabilitation requires that groundwater should be raised to levels comparable to those before regulations.

Former oxbows, paleo-meanders, and back swamps can function as natural water reservoirs, suitable for the storage of surplus water provided from adjacent streams [5]. The main objective of this research was to develop a unified management approach for water retention along the Hungarian Drava section through the implementation of a natural reservoir.

2 Study Area

The study area is located along the Lower Drava River, SW-Hungary. The floodplain of silty-sandy alluvium and fluvisols lies at 97–107 m elevation, crossed by 13 tributary streams, with 20 side channels, and 18 oxbow lakes [2]. The physical properties of soils are controlled by subdued ridge-and-swale topography [6]. The Korcsina subarea is a paleo-channel system of potential water reservoir (Fig. 1).

Fig. 1
figure 1

Location of study area and elevation map

Full size image

3 Groundwater Flow Model Setup

The MODFLOW-NWT model [7] with ModelMuse [8], which is a graphical user interface, was used to simulate 3D groundwater flows in quadratic 30 × 30 m cells. Ground penetration radar records and auger holes show high heterogeneity in the hydraulic proprieties of sediment at 0–10 m [9]. Long-term average annual precipitation and temperature are 700 mm/year and 10.5 °C, respectively. Hydraulic conductivity was measured using the fall head method. The N, S, and E boundaries were represented by general head boundary (GHB) based on the records from four observation wells. The conductance of GHB was initially set to 5 m2 days−1. The Sió-pusztai and Korcsina streams were assigned by a river package. Groundwater recharge and evapotranspiration were calculated using WetSpass-M, a spatially distributed water balance model [10].

A steady-state model was developed to calibrate the original model using groundwater levels for one auger hole and three observation wells. The calibrated values of hydraulic conductivity resulted in five zones: 300, 150, 60, 17, and 0.03 m days−1. The conductance of stream beds was 30 m2 days−1, while the conductance of the GHB boundary changed to 15 m2 days−1. The calibrated results showed good agreement (correlation coefficient: 0.93, mean error: 0.04 m, and mean absolute error: 0.05 m).

3.1 Water Balance

In the water balance of the calibrated model (Table 1), total groundwater recharge was 29,765,891 m3 days−1, i.e., 70% of the total inflow to the aquifer, while most of the aquifer outflow was directed to the river and comprised by evapotranspiration, 25,971 m3 days−1 and 2,703 m3 days−1, respectively.

Table 1 Water balance for the simulated steady-state model
Full size table

3.2 Scenarios of Water Retention Simulation for Potential Water Reservoirs

The planned reservoir (Fig. 1) was simulated by the reservoir package. Re- and discharge conditions were based on +2 m vertical water level increase. Feeding was from Korcsina channel at 2 m3 s−1 (172,800 m3 days−1) discharge to reach +2.0 m elevation in 33 days with 6,625,675 m3 water. During the recharge period, subsurface seepage tended to increase (Fig. 2a). The difference between the amount of filling and seepage gave the storage volume at the natural reservoir site (a paleo-channel). The discharge caused by subsurface seepage and less evapotranspiration showed that this reservoir system lost the stored water within 2 weeks (Fig. 2b).

Fig. 2
figure 2

Filling volume, seepage [m3], and water level of the reservoir for the recharge period (a) and water loss from the reservoir related to water level (b)

Full size image

The first 1-m increase causes soil saturation, while further filling creates a surface water body (intermittent lake). The simulated groundwater head at the end of recharge and discharge periods is shown in Fig. 3a,b. The subsurface inflow flows through the NE boundary while the groundwater levels decreased in the western part. Seepage creates a subsurface water surplus in the root zones of the surrounding area.

Fig. 3
figure 3

Simulated groundwater head at the end of recharge period on the 34th day (a) and for the discharge scenario on the 14th day (b)

Full size image

4 Conclusion

In past centuries, human activities and Drava River incision increased the drought hazard in the Drava floodplain through lowering the groundwater table. Consequently, agricultural production and wetland habitats reached a serious situation. The study investigated water retention opportunities in the desiccating lower Drava floodplain through the implementation of natural reservoirs. The Korcsina subarea was simulated using a groundwater model (MODFLOW-NWT) to assess the dynamics of a natural reservoir and groundwater interactions for both recharge and discharge scenarios for the reservoir. The outputs show the efficiency of a natural reservoir in enhancing soil saturation for the first 1-m level rise and working as an open water surface for the second meter which is useful to protect wetland habitats in the floodplain. The integrated model is envisioned to be a step toward sustainable management of the Drava floodplain.