Search

Monitoring and providing accurate land use and land cover (LULC) change information is vital for sustainable environmental planning. This study used Landsat imagery from 2002 to 2022 to create updated LULC change maps for the eThekwini Municipality. Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) were used to conduct these LULC classifications, with XGBoost achieving the highest accuracy (80.57%). The generated maps revealed a significant decrease in cropland and an increase in impervious surfaces. As such, this research established a framework for continuous LULC mapping and highlighted Landsat 9’s potential in LULC classifications. [ABSTRACT FROM AUTHOR]

More than 75% of the global land has already suffered degradation, leading to the recognition of land degradation as one of the foremost challenges society faces. This recognition stems from its profound adverse impacts on natural ecosystem functioning, biodiversity, soil productivity, and food availability. Consequently, understanding the spatial distribution of land degradation across all scales becomes imperative. This study employed land cover change and soil organic carbon (SOC) stock assessments to analyse land degradation within the eThekwini Municipality beyond the baseline period (2000–2015). Utilizing remote sensing and machine learning techniques, this research examined land degradation within the eThekwini Municipality over the period spanning 2000 to 2022. Landsat 7 (Enhanced Thematic Mapper Plus – ETM+), Landsat 8 (Operational Land Imager 1 - OLI1), and Landsat 9 (Operational Land Imager 2 - OLI2) images were employed to extract variables for both land cover change and SOC stock prediction through XGBoost, LightGBM, Random Forest (RF), and Support Vector Machine (SVM) models. Among these models, LightGBM demonstrates superior performance, achieving an overall accuracy of 80.646 in land cover predictions and 77.869 in SOC stock predictions. Analysis of land cover change within the eThekwini Municipality unveiled a shift from forests and shrubland landscapes to cropland and built-up areas. This shift results in the municipality encountering losses in SOC stock between 2015 and 2022. The model predicted that most SOC stock losses occur at the 20–50 cm depth (9.27%), in comparison to the 7.21% loss at the 0–20 cm depth. These findings underscore the pivotal role of remote sensing and machine learning in aiding policymakers to assess land degradation and implement pertinent measures to enhance the landscape. [ABSTRACT FROM AUTHOR]

Grasslands are biomes of significant fiscal, social and environmental value. Grassland or rangeland management often monitors and manages grassland productivity. Productivity is determined by various biophysical parameters, one such being grass aboveground biomass. Advancements in remote sensing have enabled near-real-time monitoring of grassland productivity. Furthermore, the increase in sophisticated machine learning algorithms has provided a powerful tool for remote sensing analytics. This study compared the performance of two neural networks, namely, Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN), in predicting dry season aboveground biomass using open-access Sentinel-2 MSI data. Sentinel-2 spectral bands and derived vegetation indices were used as input data for the two algorithms. Overall, findings in this study showed that the deep CNN outperformed the ANN in estimating aboveground biomass with an R2 of 0.83, an RMSE of 3.36 g/m2 and an RMSE% of 6.09. In comparison, the ANN produced an R2 of 0.75, an RMSE of 5.78 g/m2 and an RMSE% of 8.90. The sensitivity analysis suggested that the blue band, Green Chlorophyll Index (GCl), and Green Normalised Difference Vegetation Index (GNDVI) were the most significant for model development for both neural networks. This study can be considered a pilot study as it is one of the first to compare different neural network performances using freely available satellite data. This is useful for more rapid biomass estimation, and this study exhibits the great potential of deep learning for remote sensing applications. [ABSTRACT FROM AUTHOR]

The assessment of carbon storage in natural forests is paramount in the ongoing efforts against climate change. While traditional field-based methods for quantifying carbon storage pose challenges, recent advancements in remote sensing and machine learning offer efficient and innovative alternatives. This systematic literature review investigates the latest developments in utilising optical, radar, and light detection and ranging (LiDAR) remote sensing data, coupled with cutting-edge machine learning algorithms, to estimate carbon storage in natural forests. Non-parametric machine-learning algorithms commonly applied to multispectral datasets have emerged as prominent tools for predicting aboveground carbon storage. Nonetheless, accurately assessing forest carbon storage using remote sensing data can be arduous in regions characterised by complex terrain and diverse species where dataset noise may be pronounced. Alternatively, the adoption of freely available optical sensors with moderate resolution has showcased reliability in estimating forest carbon storage. Hence, leveraging the integration of multi-sensor data with machine learning techniques has yielded substantial improvements in the accuracy of carbon storage estimation. This study identifies the most sensitive remote sensing variables that correlate with measurable biophysical parameters, thus highlighting the pivotal role of geospatial technologies in estimating terrestrial aboveground carbon storage. The study also delineates gaps and limitations inherent in current practices, underscoring the need for further investigations in this rapidly evolving field. Through the unification of conventional methods with state-of-the-art technologies, this study contributes to the advancement of accurate and efficient carbon storage assessments. By assuming such a transformative role, this research holds substantial promise in bolstering global climate change mitigation efforts. Ultimately, the purpose of this study was to demonstrate to researchers, policy makers and practitioners the importance of embracing the combined power of remote sensing and machine learning as a tool for safeguarding our natural forests and fight against climate change. [ABSTRACT FROM AUTHOR]

Commercial forest plantations in South Africa are homogeneous monocultures of highly bred exotic species grown to deliver timber products of the best potential quality. As such, these stands are susceptible to adverse effects of biotic and abiotic factors, and therefore require intense management to mitigate these risks. A sustainable forest monitoring system that can detect real-time changes in the physiological state of these plantations is needed for timeous management intervention to reduce losses. The use of machine learning algorithms has recently become popular, with acceptable levels of success. This study explores the application of deep learning neural networks for early detection of damage caused by baboons in evergreen plantations of Pinus species. Using PlanetScope imagery (spectral band 590–860 nm), which is captured by a constellation of Dove nanosatellites, with a high temporal resolution available daily at 3 m spatial resolution, the study achieved an overall accuracy of 81.54%, with a kappa value of 0.69, using a deep neural network. In comparison, using a random-forest classifier produced 74.04% accuracy and a kappa value of 0.62. The study successfully mapped different levels of baboon damage within commercial pine forests. We provide a repeatable method for daily monitoring initiatives, and attest to the utility of higher-resolution imagery such as PlanetScope for mapping health and damage severity at the tree level. [ABSTRACT FROM AUTHOR]

Leaf area index (LAI) is an important parameter in plant ecophysiology; it can be used to quantify foliage directly and as a measure of the photosynthetic active area and, thus, the area subject to transpiration in vegetation. The aim of this paper was to review work on remote sensing methods of estimating LAI across different forest ecosystems, crops and grasslands in terms of remote sensing platforms, sensors and models. To achieve this aim, scholarly articles with the title or keywords "Leaf Area Index estimation" or "LAI estimation" were searched on Google Scholar and Web of Science with a date range between 2010 and 2020. The study's results revealed that during the last decade, the use of remote sensing to estimate and map LAI increased for crops and natural forests. However, there is still a need for more research concerning commercial forests and grasslands, as the number of studies remains low. Of the 84 studies related to forests, 60 were related to natural forests and 24 were related to commercial forests. In terms of model types, empirical models were most often used for estimating the LAI of forests, followed by physical models. [ABSTRACT FROM AUTHOR]

The detection and monitoring of lethal pathogenic funguses are important for effectively deploying suppressive measures to sudden and severe outbreaks in plantation forestry. This study successfully investigated the utility of Landsat 8 multispectral satellite imagery to map Uromycladium acacia (wattle rust) induced canopy defoliation across black wattle plantations. The framework developed for the provincial assessment of rust damage proved to be effective by using data collected from field monitoring plots over the year 2015 and 2016. Using a powerful Gradient Boosting Machine (GBM) approach, rust occurrences were mapped at accuracies of 69% for March and 72% for November when using the 2015 dataset and 77% for March and 81% for November using the 2016 dataset. Individual class accuracies for varying levels of defoliation were also evaluated. When aggregating the field and image datasets, a two-year probability map revealed the likelihood of rust defoliation across the black wattle plantation region. Overall, the study showcased the robustness and cost-effectiveness of using multispectral remote sensing methodologies for repeatable forest health monitoring in key commercial forest plantations. [ABSTRACT FROM AUTHOR]

Drought frequency is expected to increase in the coming decades due to climate change. Droughts are one of the most devastating natural disasters affecting food production, water resources and causing widespread human and animal mortalities. Sub-Saharan Africa is considered the most vulnerable region to climate change, due to low rainfall, long dry seasons and high levels of poverty. A systematic approach was used to search for published literature from 2008 to 2018 on the impacts of droughts on small-scale farmers in Sub-Saharan Africa from Ebscohost, Google Scholar and Web of Science internet search engines. To acquire relevant studies, the following keywords and Boolean operation combinations were used: GIS AND Remote Sensing OR drought OR small-scale farmers OR sub-Saharan Africa and only studies from 2008 to 2018 were selected. The impacts of droughts are far-reaching and affect the environment, societies and the economy of a country. There is limited reliable and comprehensive information on the impacts of droughts, which calls for research aimed at explaining the impacts of effect and how they can be predicted and mitigated. Improving drought preparedness and mitigation is an important precondition reducing the vulnerability of small-scale farmers and rural communities to the impacts of droughts. [ABSTRACT FROM AUTHOR]

Bracken fern is one of the major invasive plants distributed all over the world currently threatening socio-economic and ecological systems due to its ability to swiftly colonize landscapes. The study aimed at reviewing the progress and challenges in detecting and mapping of bracken fern weeds using different remote sensing techniques. Evidence from literature have revealed that traditional methods such as field surveys and modelling have been insufficient in detecting and mapping the spatial distribution of bracken fern at a regional scale. The applications of medium spatial resolution sensors have been constrained by their limited spatial, spectral and radiometric capabilities in detecting and mapping bracken fern. On the other hand, the availability of most of these data-sets free of charge, large swath width and their high temporal resolution have significantly improved remote sensing of bracken fern. The use of commercial satellite data with high resolution have also proven useful in providing fine spectral and spatial resolution capabilities that are primarily essential to offer precise and reliable data on the spatial distribution of invasive species. However, the application of these data-sets is largely restricted to smaller areas, due to high costs and huge data volumes. Studies on bracken fern classification have extensively adopted traditional classification methods such as supervised maximum likelihood classifier. In studies where traditional methods performed poorly, the combination of soft classifiers such as super resolution analysis and traditional methods of classification have shown an improvement in bracken fern classification. Finally, since high spatial resolution sensors are expensive to acquire and have small swath width, the current study recommends that future research can also consider investigating the utility of the freely available recently launched sensors with a global footprint that has the potential to provide invaluable information for repeated measurement of invasive species over time and space. [ABSTRACT FROM AUTHOR]

The objective of this study was to identify an appropriate spatial resolution for discriminating forest vegetation at subspecies level. WorldView-2 imagery was progressively resampled to coarser spatial resolutions. At a compartment level, 30 × 30-m subsets were generated across forest compartments to represent the five forest subspecies investigated in this study. From the centre of each subset, the spatial resolution of the original WorldView-2 image was resampled from 6 to 34-m, with increments of 4-m. The variance was then calculated at every resampled spatial resolution using each of the eight WorldView-2 bands. Based on the sampling theorem, the 3-m spatial resolution provided an appropriate resolution for all subspecies investigated. The WorldView-2 image was subsequently classified using the partial least squares linear discriminant analysis algorithm and the appropriate spatial resolution. An overall classification accuracy of 90% was established with an allocation disagreement of 9 and a quantity disagreement of 1. [ABSTRACT FROM AUTHOR]

Coryphodema tristis is a wood-boring insect, indigenous to South Africa, that has recently been identified as an emerging pest feeding on Eucalyptus nitens, resulting in extensive damage and economic loss. Eucalyptus plantations contributes over 9% to the total exported manufactured goods of South Africa which contributes significantly to the gross domestic product. Currently, the distribution extent of the Coryphodema tristis is unknown and estimated to infest Eucalyptus nitens compartments from less than 1% to nearly 80%, which is certainly a concern for the forestry sector related to the quantity and quality of yield produced. Therefore, the study sought to model the probability of occurrence of Coryphodema tristis on Eucalyptus nitens plantations in Mpumalanga, South Africa, using data from the Sentinel-2 multispectral instrument (MSI). Traditional field surveys were carried out through mass trapping in all compartments (n = 878) of Eucalyptus nitens plantations. Only 371 Eucalyptus nitens compartments were positively identified as infested and were used to generate the Coryphodema tristis presence data. Presence data and spectral features from the area were analysed using the Maxent algorithm. Model performance was evaluated using the receiver operating characteristics (ROC) curve showing the area under the curve (AUC) and True Skill Statistic (TSS) while the performance of predictors was analysed with the jack-knife. Validation of results were conducted using the test data. Using only the occurrence data and Sentinel-2 bands and derived vegetation indices, the Maxent model provided successful results, exhibiting an area under the curve (AUC) of 0.890. The Photosynthetic vigour ratio, Band 5 (Red edge 1), Band 4 (Red), Green NDVI hyper, Band 3 (Green) and Band 12 (SWIR 2) were identified as the most influential predictor variables. Results of this study suggest that remotely sensed derived vegetation indices from cost-effective platforms could play a crucial role in supporting forest pest management strategies and infestation control. [ABSTRACT FROM AUTHOR]

The invasive weed Solanum mauritianum (bugweed) has infested large areas of plantation forests in KwaZulu-Natal, South Africa. Bugweed often forms dense infestations and rapidly capitalises on available natural resources hindering the production of forest resources. Precise assessment of bugweed canopy cover, especially at low abundance cover, is essential to an effective weed management strategy. In this study, the utility of AISA Eagle airborne hyperspectral data (393–994 nm) with the new generation Worldview-2 multispectral sensor (427–908 nm) was compared to detect the abundance of bugweed cover within the Hodgsons Sappi forest plantation. Using sparse partial least squares discriminant analysis (SPLS-DA), the best detection results were obtained when performing discrimination using the remotely sensing images combined with LiDAR. Overall classification accuracies subsequently improved by 10% and 11.67% for AISA and Worldview-2 respectively, with improved detection accuracies for bugweed cover densities as low as 5%. The incorporation of LiDAR worked well within the SPLS-DA framework for detecting the abundance of bugweed cover using remotely sensed data. In addition, the algorithm performed simultaneous dimension reduction and variable selection successfully whereby wavelengths in the visible (393–670 nm) and red-edge regions (725–734 nm) of the spectrum were the most effective. [ABSTRACT FROM AUTHOR]

• The majority of KwaZulu-Natal's SOC stocks were found to be distributed between 60–200 cm. • Forested areas in KwaZulu-Natal had the greatest SOC concentration per unit Area. • Rainfall, clay content, temperature, elevation, and RVI were found to be key determinants of SOC prediction at different depths. • DNN model performed well in topsoil (0–30 cm), but less accurate with deeper depth. The preservation and augmentation of soil organic carbon (SOC) stocks is critical to designing climate change mitigation strategies and alleviating global warming. However, due to the susceptibility of SOC stocks to environmental and topo-climatic variability and changes, it is essential to obtain a comprehensive understanding of the state of current SOC stocks both spatially and vertically. Consequently, to effectively assess SOC storage and sequestration capacity, precise evaluations at multiple soil depths are required. Hence, this study implemented an advanced Deep Neural Network (DNN) model incorporating Sentinel-1 Synthetic Aperture Radar (SAR) data, topo-climatic features, and soil physical properties to predict SOC stocks at multiple depths (0–30 cm, 30–60 cm, 60–100 cm, and 100–200 cm) across diverse land-use categories in the KwaZulu-Natal province, South Africa. There was a general decline in the accuracy of the DNN model's prediction with increasing soil depth, with the root mean square error (RMSE) ranging from 8.34 t/h to 11.97 t/h for the four depths. These findings imply that the link between environmental covariates and SOC stocks weakens with soil depth. Additionally, distinct factors driving SOC stocks were discovered in both topsoil and deep-soil, with vegetation having the strongest effect in topsoil, and topo-climate factors and soil physical properties becoming more important as depth increases. This underscores the importance of incorporating depth-related soil properties in SOC modelling. Grasslands had the largest SOC stocks, while commercial forests have the highest SOC sequestration rates per unit area. This study offers valuable insights to policymakers and provides a basis for devising regional management strategies that can be used to effectively mitigate climate change. [ABSTRACT FROM AUTHOR]