Semi-Supervised Learning for Grain Size Distribution Interpolation

Kobs, Konstantin; Schäfer, Christian; Steininger, Michael; Krause, Anna; Baumhauer, Roland; Paeth, Heiko; Hotho, Andreas

doi:10.1007/978-3-030-68780-9_4

Cited by 2 publications

(1 citation statement)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For further investigations, new interpolation methods can be developed to describe in-depth (forest-)soil hydrology by a combination of state-of-the-art land use and soil maps, observations, and advanced pedotransfer functions. For example, there exist several approaches to increase the spatial resolution, like Random Forests or Neural Networks (Taghizadeh-Mehrjardi et al, 2020;Kobs et al, 2021) that could also be applied to other regions. A more differentiated land use map, e.g., by tree species, would also be possible according to the current state of knowledge in remote sensing and should be prepared for future modeling.…”

Section: Outlookmentioning

confidence: 99%

Multidimensional hydrological modeling of a forested catchment in a German low mountain range using a modular runoff and water balance model

Schäfer¹,

Fäth²,

Kneisel³

et al. 2023

Front. For. Glob. Change

View full text Add to dashboard Cite

Sufficient plant-available water is one of the most important requirements for vital, stable, and well-growing forest stands. In the face of climate change, there are various approaches to derive recommendations considering tree species selection based on plant-available water provided by measurements or simulations. Owing to the small-parcel management of Central European forests as well as small-spatial variation of soil and stand properties, in situ data collection for individual forest stands of large areas is not feasible, considering time and cost effort. This problem can be addressed using physically based modeling, aiming to numerically simulate the water balance. In this study, we parameterized, calibrated, and verified the hydrological multidimensional WaSiM-ETH model to assess the water balance at a spatial resolution of 30 m in a German forested catchment area (136.4 km2) for the period 2000–2021 using selected in situ data, remote sensing products, and total runoff. Based on the model output, drought-sensitive parameters, such as the difference between potential and effective stand transpiration (Tdiff) and the water balance, were deduced from the model, analyzed, and evaluated. Results show that the modeled evapotranspiration (ET) correlated significantly (R2 = 0.80) with the estimated ET using MODIS data (MOD16A2GFv006). Compared with observed daily, monthly, and annual runoff data, the model shows a good performance (R2: 0.70|0.77|0.73; Kling–Gupta efficiency: 0.59|0.62|0.83; volumetric efficiency: 0.52|0.60|0.83). The comparison with in situ data from a forest monitoring plot, established at the end of 2020, indicated good agreement between observed and simulated interception and soil water content. According to our results, WaSiM-ETH is a potential supplement for forest management, owing to its multidimensionality and the ability to model soil water balance for large areas at comparable high spatial resolution. The outputs offer, compared to non-distributed models (like LWF-Brook90), spatial differentiability, which is important for small-scale parceled forests, regarding stand structure and soil properties. Due to the spatial component offered, additional verification possibilities are feasible allowing a reliable and profound verification of the model and its parameterization.

show abstract

Section: Outlookmentioning

confidence: 99%