The Impact of Soil Moisture on Groundwater Level Forecasting Using Deep Neural Networks: Evidence from Brandenburg, Germany
Bibliographic record
Abstract
One major challenge in reliable groundwater level forecasting is to correctly account for the amount and rate of precipitation percolating through the unsaturated zone prior to reaching the aquifer. Especially under a changing climate already impacting weather and climate extremes globally, increased frequency of heatwaves, heavy precipitation, and drought periods will have significant impact on recharge patterns through soil hydraulic properties and unsaturated zone dynamics. However, as soon as groundwater predictions concern long-term environmental changes, extrapolations beyond the short-term often lack to fully account for increased frequency of extreme events under climate change. Consequently, estimates and forecasts overlook the actual impacts of weather extremes, particularly imprinting themselves in changes in the hydraulic connection between groundwater and soil surface.We used weekly groundwater level data (1990 – 2024) from over a hundred measuring wells, well distributed over the federal state of Brandenburg, Germany, to train a deep neural network, that is able to predict groundwater level development under the impacts of climate change. To account for the soil hydraulic properties, we included soil moisture from different depths as a proxy for the amount and timing of water percolating through the vadose zone.We show that purely climatic inputs, such as air temperature and precipitation are not sufficient to explain regional groundwater level development, as suggested by previous studies. Instead, including soil moisture turns out be the factor with the highest impact (feature importance) on the entire regional model, increasing the explained variance for most sites, while being able to reduce the model error constantly (RSME). Our findings demonstrate that future predictions of groundwater level can be enhanced by integrating the effects of climate on soil moisture into predictive models.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".