The need for the software is based on being able to make a statement as to whether the operation of a Pumped Hydropower Storage (PHS) facility in a former open-pit lignite mine can have a negative impact on the water quality in the lower reservoir and associated aquifers. The research question arises since flooded lignite mines are often associated with acidification and/or increased sulphate and metal concentrations. Thus, the software aims at modelling geochemical processes during the PHS operation in open-pit lignite mines.
The reaction path modelling framework comprises a Python framework for data management and a solver for geochemical reactions (phreeqc/phreeqpy; Parkhurst and Appelo, 2013; Müller, 2011). The software is based on a conceptual geochemical model that includes the main geochemical processes that are expected to influence the hydrochemistry. It integrates different non-dimensional batch reactors, each representing the water composition of the reservoirs, and water sources or sinks in the PHS system (groundwater, rainwater, surface run-off, mine dump water). These waters are cyclically mixed with ratios deducted from flow rates and time-dependent influxes of a hypothetical PHS system.
The water influxes have different chemical compositions based on the geochemical scenarios defined with the input data. An instant flooding of the mine with scenario-specific mixing ratios of rainwater, groundwater and mine dump water is simulated to provide an initial solution in the LR for the PHS operation. For the simulation of the PHS operation, the water volume of the UR is extracted from the LR and equilibrated with atmospheric partial pressures of oxygen and carbon dioxide to represent the water composition after pumping. The water composition evolving at the reservoir-mine dump interface layer is simulated by a kinetically controlled reaction of pyrite (Williamson and Rimstid, 1994) and calcite (Plummer, 1978) with the LR water. During the PHS discharge cycle, water flows into the adjacent mine dump sediments due to the increasing hydraulic head gradient in the LR compared to the surrounding groundwater aquifers. Water from the LR is mixed with rainwater, groundwater, surface run-off, and water from the reservoir-mine dump interface layer according to the water volumes that enter the reservoir during the respective cycle. Finally, the new water composition in the LR is mixed with the water from the UR to simulate the PHS discharge into the LR. Apart from gas exchange, evaporation and precipitation, no reactions are simulated for the water in the UR, as the reservoir is assumed to be artificially sealed.
Pump and discharge cycles are simulated until the pH and sulfate concentrations in the LR do not change by more than 1 x 10-4 and 1 x 10-5 mol kgw-1 within two consecutive PHS cycles, respectively. Otherwise, the simulation is terminated after 7,300 PHS cycles, representing 20 years of operation with a duration of one day per cycle. Input parameter ranges can cover a wide range of potential hydrogeochemical scenarios. In the software provided with this manual, a small range of generic data is defined as input to limit the simulation time and data output. However, the input can be modified to simulate a broader range of geochemical scenarios as described in the associated data description file.
This dataset contains data of measurements at several inland water bodies in the region around Karlsruhe, Germany, including spectral data and mainly chlorophyll a values. The measurements are motivated by the idea to link remote sensing data and in-situ acquired data for an estimation of selected water parameters solely based on spectral data with data-driven machine learning approaches. To address this topic, we monitored the selected inland water bodies with a spectrometer as a remote sensing device and two in-situ measurement devices to collect among others the chlorophyll a values during the summers of the years 2018 and 2019. The eleven selected water bodies are characterized by a relatively small size and a variety of different chlorophyll a concentrations.
In sum, the dataset of the year 2018 consists of 1,305 datapoints and the dataset of the year 2019 includes 2,830 datapoints. Each datapoint is defined by the spectral data in the range of 389 nm to 910 nm and all in-situ measured reference values such as the chlorophyll a concentration of the respective year. Depending on the applied in-situ measurement device, the chlorophyll a concentration or additional water parameters were either measured directly on-site or taken as water samples and analyzed in the laboratory.
We provide the data in nearly raw format including information about the water bodies. The data are organized along the year of the measurements. In addition, exemplary scripts for reading and processing the data are included.