Gifty Attiah

Lake surface temperature (LST) and NDTI-based optical turbidity are important indicators of water quality and ecosystem dynamics in northern lakes, where climate warming and river discharge strongly influence thermal structure and sediment transport. This study used multi-sensor satellite observations processed in Google Earth Engine to examine long-term (1984–2024) variability in LST and NDTI-based optical turbidity within the Slave River–Great Slave Lake system, Northwest Territories, Canada, using Landsat, MODIS-Aqua and Sentinel-2 imagery. Landsat (1984–2024) and MODIS Aqua (2002–2024) indicated open-water warming of approximately +0.06 and +0.07 °C yr⁻¹, respectively; only the Landsat trend was statistically significant (p = 0.038 versus p = 0.073). The Normalized Difference Turbidity Index (NDTI), used as a relative optical proxy for turbidity, exhibited strong interannual variability that co-varied primarily with Slave River discharge (MODIS Aqua r = 0.91, n = 23, t = 9.80; Landsat 1984–2024 r = 0.72, n = 41, t = 6.41; both p < 0.0001). Principal Component Analysis revealed discharge and NDTI-based optical turbidity as the dominant modes of variability, explaining 65% of the total variance, while LST loaded mainly on a separate component, accounting for a further 32%. The 2020 high-flow event produced the most extensive NDTI-based optical turbidity episode within the 40-year study period, with elevated NDTI consistently detected across all satellite platforms, followed by a marked decline through 2024. Monthly analysis revealed a discharge-dependent NDTI-based optical turbidity lag, with peak NDTI-based optical turbidity occurring concurrently with or within one month of peak discharge during high-flow years but delayed by up to 3–4 months during low-flow years. Spatial analyses further demonstrated persistent NDTI-based optical turbidity intensification near the Slave River mouth and the southwestern basin. These findings indicate that although LST shows a positive but sensor-dependent warming tendency, NDTI variability within the Slave River Plume Zone (SRPZ) was primarily associated with episodic high-discharge conditions consistent with sediment transport processes dominating optical water quality. The study further highlights the value of multi-sensor satellite observations for long-term environmental monitoring in remote subarctic lake systems.

1. Urban heat anomalies, driven by land cover transformation and human activity, are an increasingly urgent challenge in rapidly urbanizing regions worldwide (Marchiori et al., 2012; Nasar-U-Minall...

Turbidity is a crucial indicator of water quality and ecosystem health in large northern lakes. However, logistical reasons, high costs and the spatial complexity of lakes result in a scarcity of observations, hindering long-term monitoring efforts. In this study, we develop a dual-threshold Random Forest (RF) model optimized for low- and high-turbidity regimes to retrieve turbidity from Moderate Resolution Imaging Spectroradiometer (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS) surface reflectance and band ratios over Great Slave Lake (GSL) from 2002 to 2024. Our objectives were to create a turbidity dataset from MODIS and VIIRS, assess sensor performance against in-situ measurements and among sensors, and investigate the spatial, seasonal, and interannual dynamics of turbidity. The trained RF models performed robustly across both sensors and turbid regimes: R² = 0.80 - 0.83, r = 0.89-0.91, IA = 0.94-0.95, MBE within ±0.30 NTU, with RMSE 2.3-2.8 NTU (low-turbidity areas) and 11.4 -14.1 NTU (high-turbidity areas). VIIRS consistently slightly outperformed MODIS, with lower RMSE and higher R², and has consistently higher correlations with input ratios. We observe a significant increasing trend of turbidity across the lake up to ~2 NTU/yr. Both sensors identified persistent high-turbidity zones in GSL’s southeastern region and nearshore areas, associated with river inflow from the Slave River, shallow bathymetry, and wind-driven resuspension, with plume extent expanded during high-inflow years such as 2020 and 2022. Collectively, these results provide the first harmonized continuous MODIS-VIIRS turbidity record for GSL (https://doi.org/10.20383/103.01563), quantify sensor-specific performance, and link satellite retrievals to primary limnological drivers.

The Canadian Lake Ice Model has been adapted and distributed at a high spatial resolution (∼50 m) to simulate Lake Ice Thickness (LIT) and phenology across small to medium‐sized lakes in this study. The model was applied to simulate the spatial variability of LIT, Ice Cover Duration (ICD), and Lake Ice Phenology across 500 predominantly small to medium‐sized lakes in the North Slave Region (NSR) of the Northwest Territories (NWT), Canada, from 1984 to 2022. The model utilizes a 30 m grid lake surface temperature (LST) data set derived from the North Slave LST data set, along with climate inputs from the European Center for Medium‐Range Weather Forecasts Reanalysis v5 (ERA5), including wind speed, relative humidity, snowfall, and cloud cover. These inputs provide surface fluxes to the model, driving its unsteady heat equation to produce daily LIT, annual ICD, and annual freeze‐up and break‐up dates on a ∼50 m grid resolution. Validation against in situ measurements shows a root mean square deviation of 2.7–7 cm for LIT. Trend analysis revealed a significant decline in LIT (−0.26 cm/year to −0.10 cm/year) and ICD (−0.40 days/year to −0.15 days/year). The highest rates of LIT decline are observed in the early and later months of ice formation. Freeze‐up timings are primarily influenced by depth, while geographic factors such as latitude and elevation affect break‐up timing in the NSR. This distributed modeling approach provides a robust assessment of lake ice variability and trends under changing climate and weather conditions.

Remote sensing data, particularly from Sentinel-2 satellites, offers valuable insights into surface mining activities. This study evaluates the effectiveness of Sentinel-2 imagery in detecting and monitoring sand and gravel extraction sites across five mining areas in Schleswig-Flensburg, Germany, from 2015 to 2019. Three machine learning algorithms—Random Forest (RF), Support Vector Machines (SVM), and Artificial Neural Networks (ANN)—were compared to determine the most accurate classification approach. Models were trained using three different training data scenarios to assess their performance. While RF and SVM demonstrated greater robustness than ANN, SVM outperformed the others when validated against ground truth data. Consequently, an optimized SVM-based classification model was developed and implemented in R to analyze temporal changes in the study areas over five years. The findings highlight the potential of integrating Sentinel-2 imagery with machine learning techniques for accurate and efficient monitoring of surface mining activities, offering a scalable approach for environmental management and land-use planning.

The Canadian Lake Ice Model has been distributed and adapted to run at a fine spatial resolution (~50 m) for simulating lake ice thickness and phenology on small to medium-sized lakes for this study. The model’s capabilities are extended to simulate the spatial variability of Lake ice thickness (LIT), Ice Cover Duration (ICD), and Lake Ice Phenology (LIP) across 500 predominantly small to medium-sized lakes in the North Slave Region (NSR) of the Northwest Territories (NWT), Canada, from 1984 to 2022. The model utilizes 30 m grid lake surface temperature (LST) data derived from the North Slave LST dataset, along with climate inputs from the European Centre for Medium-Range Weather Forecasts Reanalysis v5 and ECMWF (ERA5), including wind speed, mean air temperature, relative humidity, snow depth, and cloud cover. These inputs provide surface fluxes to the model, driving its unsteady heat equation to produce daily LIT, annual ICD, and annual freeze-up and break-up dates on a ~50 m grid. Validation against in-situ measurements showed a root mean square deviation of 2.7 cm to 7 cm for LIT. Trend analysis revealed a significant decline in LIT (-0.26 cm/year to -0.10 cm/year) and ICD (-0.40 day/year to -0.15 day/year) over the study period. The findings highlight the sensitivity of LIT and freeze-up dates to lake morphometry. In contrast, ice break-up dates are primarily influenced by geographic factors such as latitude and elevation. The distributed model comprehensively assesses lake ice variability and trends across 500 lakes under changing climate and weather conditions.

Lake surface temperature (LST) is an important attribute that highlights regional weather and climate variability and trends. The spatial resolution and thermal sensors on Landsat platforms provide the capability of monitoring the temporal and spatial distribution of lake surface temperature on small- to medium-sized lakes. In this study, a retrieval algorithm was applied to the thermal bands of Landsat archives to generate a LST dataset (North Slave LST dataset) for 535 lakes in the North Slave Region (NSR) of the Northwest Territories (NWT), Canada, for the period of 1984 to 2021. North Slave LST was retrieved from Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM+), and Landsat 8 Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS); however, most of the dataset was created from the thermal bands of Landsat 5 (43 %) due to its longevity (1984–2013). Cloud masks were applied to Landsat images to eliminate cloud cover. In addition, a 100 m inward buffer was applied to lakes to prevent pixel mixing with shorelines. To evaluate the algorithm applied, retrieved LST was compared with in situ data and Moderate Resolution Imaging Spectroradiometer (MODIS) LST observations. A good agreement was observed between in situ observations and North Slave LST, with a mean bias of 0.12 ^∘C and a root mean squared deviation (RMSD) of 1.7 ^∘C. The North Slave LST dataset contains more available data for warmer months (May to September; 57.3 %) compared to colder months (October to April). The average number of images per year for each lake across the NSR ranged from 20 to 45. The North Slave LST dataset, available at https://doi.org/10.5683/SP3/J4GMC2 (Attiah et al., 2022), will provide communities, scientists, and stakeholders with spatial and temporal changing temperature trends on lakes for the past 38 years.

Lake surface temperature (LST) is an important attribute that highlights regional weather and climate variability and trends. The spatial resolution and thermal sensors on Landsat platforms provide the capability of monitoring the temporal and spatial distribution of lake surface temperature on small to medium size lakes. In this study, a retrieval algorithm was applied to the thermal bands of Landsat archives to generate a LST dataset (North Slave LST dataset) for 535 lakes in the North Slave Region (NSR) of the Northwest Territories (NWT), Canada for the period of 1984 to 2021. North Slave LST was retrieved from Landsat-5 TM, Landsat-7 ETM+ and Landsat-8 OLI/TIRS, however majority of the dataset were created from the thermal bands of Landsat-5 (43 %) due to its longevity (1984–2013). Cloud masks were applied to Landsat images to eliminate cloud cover. In addition, a 100-meter inward buffer was applied to lakes to prevent pixel mixing with shorelines. To evaluate the algorithm applied, retrieved LST was compared with in-situ data and Moderate Resolution Imaging Spectroradiometer (MODIS) LST observations. A good agreement was observed between in-situ observations and North Slave LST derived in this study with a mean bias of 0.12 °C and an RMSD of 1.7 °C. The North Slave LST dataset contains more available data from warmer months (May to September), covering 57.3 % in comparison to colder months (October to April). Average number of images per year for each lake across the NSR ranged from 20 to 45. The North Slave LST dataset will provide communities, scientists and stakeholders with spatial and temporal changing trends of temperature on lakes for the past 38 years.