River temperature research and practice: Recent challenges and emerging opportunities for managing thermal habitat conditions in stream ecosystems.

. Water temperature in rivers is a crucial environmental factor with the ability to alter hydro-ecological as well as socio-economic conditions within a catchment. The development of modelling concepts for predicting river water temperature is and will be essential for effective integrated water management and the development of adaptation strategies to future global changes (e.g. climate change). This study tests the performance of six different machine-learning models: step-wise linear regression, random forest, eXtreme Gradient Boosting (XGBoost), feed-forward neural networks (FNNs), and two types of recurrent neural networks (RNNs). All models are applied using different data inputs for daily water temperature prediction in 10 Austrian catchments ranging from 200 to 96 000 km 2 and exhibiting a wide range of physiographic characteristics. The evaluated input data sets include combinations of daily means of air temperature, runoff, precipitation and global radiation. Bayesian optimization is applied to optimize the hyperparameters of all applied machine-learning models. To make the results comparable to previous studies, two widely used benchmark models are applied additionally: linear regression and air2stream. With a mean root mean squared error (RMSE) of 0.55  ∘ C, the tested models could significantly improve water temperature prediction compared to linear regression (1.55  ∘ C) and air2stream (0.98  ∘ C). In general, the results show a very similar performance of the tested machine-learning models, with a median RMSE difference of 0.08  ∘ C between the models. From the six tested machine-learning models both FNNs and XGBoost performed best in 4 of the 10 catchments. RNNs are the best-performing models in the largest catchment, indicating that RNNs mainly perform well when processes with long-term dependencies are important. Furthermore, a wide range of performance was observed for different hyperparameter sets for the tested models, showing the importance of hyperparameter optimization. Especially the FNN model results showed an extremely large RMSE standard deviation of 1.60  ∘ C due to the chosen hyperparameters. This study evaluates different sets of input variables, machine-learning models and training characteristics for daily stream water temperature prediction, acting as a basis for future development of regional multi-catchment water temperature prediction models. All preprocessing steps and models are implemented in the open-source R package wateRtemp to provide easy access to these modelling approaches and facilitate further research.

/pdf/machine-learning-methods-for-stream-water-temperature-423bmaxspo.pdf

Machine learning methods for stream water temperature prediction

Mountainous regions are more sensitive to climatic condition changes and are susceptible to recent increases in temperature. Due to urbanization and land use/land cover (LULC) issues, Cameron Highlands has been impacted by rising land surface temperature (LST) variation. Thus, this study was carried out to explore the impact of the LULC change on LST in the Cameron Highlands from 2009 to 2019 using remote sensing images acquired from Landsat 7 ETM+, Landsat 8 Operational Land Imager (OLI/TIRS), and Moderate Resolution Imaging Spectroradiometer (MODIS) 11A Thermal sensors. A split-window algorithm was applied to Landsat 8 images (2013–2019) to derive the LST. Air temperature data of the study area were also obtained to cross-validate data sources. Based on the validation results, the accuracy of LULC and LST outputs were more than 94.6% and 80.0%, respectively. The results show that the current trend of urban growth continues at a rate of 0.16% per year, and the area experienced an LST increase of 2 °C between 2009 and 2019. This study is crucial for land planners and environmentalists to understand the impacts of LULC change on LST and to propose appropriate policy measures to control development in Cameron Highlands.

https://www.mdpi.com/2073-445X/9/10/372/pdf

Land Use/Land Cover Changes and the Relationship with Land Surface Temperature Using Landsat and MODIS Imageries in Cameron Highlands, Malaysia

The Cameron Highlands has experienced multiple land encroachment activities and repeated deforestation, leading to extensive land-use and land-cover change (LULCC) during the past six decades. This study aims to determine the LULCC against topography in Cameron Highlands between 2009 and 2019 by using geospatial techniques to analyze Landsat 7 (ETM+) and 8 (OLI/TIRS), ASTER GDEM and MODIS imaging sensors. The results showed a decline of 35.98 km2 in primary forests over ten years across the Cameron Highlands, while agricultural lands and urban areas flourished by a rise of 51.61 km2 and 11.00 km2 respectively. It can be noted that the elevation most affected is between 1000 and 1500 m, across all classes. Further results showed the expansion of both agriculture and urban development onto slopes above 35°, leading to an instability of soil structure. In a comparison of the base years of 2009 with 2019, mean LST results have shown temperatures rising by 7.5°C, while an average between 3 and 4°C across the region is recorded. The results obtained provide new information for government bodies and land planners to coordinate their actions without further jeopardizing the environment of the Cameron Highlands.

/pdf/evaluating-the-impacts-of-land-use-land-cover-changes-across-3vufhjigf6.pdf

Evaluating the impacts of land use/land cover changes across topography against land surface temperature in Cameron Highlands.

Despite efforts to understand the hydrologic impact of hydropower dams, their influence on downstream river temperatures has gone unnoticed in data limited regions. Using 30 years of Landsat thermal infrared observations (1988-2018), we identified a relationship between dry season water temperature cooling trends and dam development in the 3S Basin, a major tributary of the Mekong River. Within a year of the beginning of operations of major dams in the 3S River Basin, rapid decreases in annual average dry season river temperature were observed ranging between 0.7° C and 2° C. Furthermore, in-situ water temperature observations confirmed decreasing river temperature for two major dam development events. Evidence was found that the 3S outflow has been cooling the Mekong River downstream of the confluence, by as much as 0.8° C in recent years. Our findings are critically important for understanding how fish and aquatic ecosystems will behave in the future as more hydropower dams are built in the Mekong River Basin.

https://iopscience.iop.org/article/10.1088/1748-9326/ab763d/pdf

Hydropower's hidden transformation of rivers in the Mekong

Water temperature regulates many processes in lakes; therefore, evaluating it is essential to understand its ecological status and functioning, and to comprehend the impact of climate change. Although few studies assessed the accuracy of individual sensors in estimating lake-surface-water temperature (LSWT), comparative analysis considering different sensors is still needed. This study evaluated the performance of two thermal sensors, MODIS and Landsat 7 ETM+, and used Landsat methods to estimate the SWT of a large subtropical lake. MODIS products MOD11 LST and MOD28 SST were used for comparison. For the Landsat images, the radiative transfer equation (RTE), using NASA’s Atmospheric Correction Parameter Calculator (AtmCorr) parameters, was compared with the single-channel algorithm in different approaches. Our results showed that MOD11 obtained the highest accuracy (RMSE of 1.05 ° C), and is the recommended product for LSWT studies. For Landsat-derived SWT, AtmCorr obtained the highest accuracy (RMSE of 1.07 ° C) and is the recommended method for small lakes. Sensitivity analysis showed that Landsat-derived LSWT using the RTE is very sensitive to atmospheric parameters and emissivity. A discussion of the main error sources was conducted. We recommend that similar tests be applied for Landsat imagery on different lakes, further studies on algorithms to correct the cool-skin effect in inland waters, and tests of different emissivity values to verify if it can compensate for this effect, in an effort to improve the accuracy of these estimates.

https://www.mdpi.com/2073-4441/11/1/168/pdf

Comparison of Methods to Estimate Lake-Surface-Water Temperature Using Landsat 7 ETM+ and MODIS Imagery: Case Study of a Large Shallow Subtropical Lake in Southern Brazil

Derivation of consistent, continuous daily river temperature data series by combining remote sensing and water temperature models

Hydrodynamic models based on the RANS equation are well-established tools to simulate three-dimensional free surface flows in large aquatic ecosystems. However, when the ratio of vertical to horizontal motion scales is not small, a non-hydrostatic approximation is needed to represent these processes accurately. Increasing efforts have been made to improve the efficiency of non-hydrostatic hydrodynamic models, but these improvements require higher implementation and computational costs. In this paper, we proposed a novel free-surface boundary condition based on a fictional sublayer at the free-surface (FSFS). We applied the FSFS approach at a finite difference numerical discretization with a fractional step framework, which uses a Neumann type of boundary condition to apply a hydrostatic relation in the top layer. To evaluate the model performance, we compared the Classic Boundary Condition Approach (CBA) and the FSFS approach using two numerical experiments. The experiments tested the model’s phase error, capability in solving wave celerity and simulate non-linear wave propagation under different vertical resolution scenarios. Our results showed that the FSFS approach had a lower phase error (2 to 5 times smaller) than CBA with a little additional computational cost (ca. 7% higher). Moreover, it can better represent wave celerity and frequency dispersion with 2 times fewer layers and low mean computational cost (CBA δ t = 2.62 s and FSFS δ t = 1.22 s).

Improvement of non-hydrostatic hydrodynamic solution using a novel free-surface boundary condition

In three-dimensional simulations of free-surface flow where the vertical velocities are relevant, such as in lakes, estuaries, reservoirs, and coastal zones, a nonhydrostatic hydrodynamic approach may be necessary. Although the nonhydrostatic hydrodynamic approach improves the physical representation of pressure, acceleration and velocity fields, it is not free of numerical diffusion. This numerical issue stems from the numerical solution employed in the advection and diffusion terms of the Reynolds-averaged Navier–Stokes (RANS) and solute transport equations. The combined use of high-resolution schemes in coupled nonhydrostatic hydrodynamic and solute transport models is a promising alternative to minimize these numerical issues and determine the relationship between numerical diffusion in the two solutions. We evaluated the numerical diffusion in three numerical experiments, for different purposes: The first two experiments evaluated the potential for reducing numerical diffusion in a nonhydrostatic hydrodynamic solution, by applying a quadratic interpolator over a Bilinear, applied in the Eulerian–Lagrangian method (ELM) step-ii interpolation, and the capability of representing the propagation of complex waves. The third experiment evaluated the effect on numerical diffusion of using flux-limiter schemes over a first-order Upwind in solute transport solution, combined with the interpolation methods applied in a coupled hydrodynamic and solute transport model. The high-resolution methods were able to substantially reduce the numerical diffusion in a solute transport problem. This exercise showed that the numerical diffusion of a nonhydrostatic hydrodynamic solution has a major influence on the ability of the model to simulate stratified internal waves, indicating that high-resolution methods must be implemented in the numerical solution to properly simulate real situations.

https://www.mdpi.com/2073-4441/11/11/2288/pdf

Augusto Hugo Farias da Cunha

Papers

Comparison of Methods to Estimate Lake-Surface-Water Temperature Using Landsat 7 ETM+ and MODIS Imagery: Case Study of a Large Shallow Subtropical Lake in Southern Brazil

Derivation of consistent, continuous daily river temperature data series by combining remote sensing and water temperature models

Improvement of non-hydrostatic hydrodynamic solution using a novel free-surface boundary condition

Combined use of high-resolution numerical schemes to reduce numerical diffusion in coupled nonhydrostatic hydrodynamic and solute transport model.

Applying a water temperature model to a river and calibrating it with downscaled remote sensing data