Timothy Dube

Intra-and-Inter Species Biomass Prediction in a Plantation Forest: Testing the Utility of High Spatial Resolution Spaceborne Multispectral RapidEye Sensor and Advanced Machine Learning Algorithms

The quantification of aboveground biomass using remote sensing is critical for better understanding the role of forests in carbon sequestration and for informed sustainable management. Although remote sensing techniques have been proven useful in assessing forest biomass in general, more is required to investigate their capabilities in predicting intra-and-inter species biomass which are mainly characterised by non-linear relationships. In this study, we tested two machine learning algorithms, Stochastic Gradient Boosting (SGB) and Random Forest (RF) regression trees to predict intra-and-inter species biomass using high resolution RapidEye reflectance bands as well as the derived vegetation indices in a commercial plantation. The results showed that the SGB algorithm yielded the best performance for intra-and-inter species biomass prediction; using all the predictor variables as well as based on the most important selected variables. For example using the most important variables the algorithm produced an R2 of 0.80 and RMSE of 16.93 t·ha−1 for E. grandis; R2 of 0.79, RMSE of 17.27 t·ha−1 for P. taeda and R2 of 0.61, RMSE of 43.39 t·ha−1 for the combined species data sets. Comparatively, RF yielded plausible results only for E. dunii (R2 of 0.79; RMSE of 7.18 t·ha−1). We demonstrated that although the two statistical methods were able to predict biomass accurately, RF produced weaker results as compared to SGB when applied to combined species dataset. The result underscores the relevance of stochastic models in predicting biomass drawn from different species and genera using the new generation high resolution RapidEye sensor with strategically positioned bands.

Detection of land cover changes around Lake Mutirikwi, Zimbabwe, based on traditional remote sensing image classification techniques

Land cover changes around Lake Mutirikwi in 1984–2011 were mapped from Landsat images using traditional image classification methods including the maximum likelihood classifier algorithm. The possibility of mapping the coverage and abundance of surface floating aquatic weeds was also tested. Landsat images from 1984, 1995, 2001 and 2011 were used to compute a normalised difference vegetation index (NDVI), which was then used as a proxy for indicating areas infested by surface floating aquatic weeds. Forest and shrubs covered 310.8 km2 in 1984, but had deteriorated by 24.87% to 77.3 km2 in 2011, while the area under cultivation increased by 51.44% between 1984 and 2011. Runoff from surrounding farms could be responsible for washing soil nutrients into Lake Mutirikwi, enriching its water. A large aggregation of surface floating aquatic weeds concentrated upstream along tributaries of Lake Mutirikwi, mainly the Mucheke which received sewage from Masvingo town, with less coverage in the central parts of the lake. Vegetation indices such as NDVI proved successful as a proxy for mapping the coverage of surface floating aquatic weeds in Lake Mutirikwi in space and time.

Employing ground and satellite-based QuickBird data and random forest to discriminate five tree species in a Southern African Woodland

With the emergence of very high spatial and spectral resolution data set, the resolution gap that existed between remote-sensing data set and aerial photographs has decreased. The decrease in resolution gap has allowed accurate discrimination of different tree species. In this study, discrimination of indigenous tree species (n = 5) was carried out using ground based hyperspectral data resampled to QuickBird bands and the actual QuickBird imagery for the area around Palapye, Botswana. The purpose of the study was to compare the accuracies of resampled hyperspectral data (resampled to QuickBird sensors) with the actual image (QuickBird image) in discriminating between the indigenous tree species. We performed Random Forest (RF) using canopy reflectance taking from ground-based hyperspectral sensor and the reflectance delineated regions of the tree species. The overall accuracies for classifying the five tree species was 79.86 and 88.78% for both the resampled and actual image, respectively. We observed that resampled data set can be upscale to actual image with the same or even greater level of accuracy. We therefore conclude that high spectral and spatial resolution data set has substantial potential for tree species discrimination in savannah environments.

Predicting Eucalyptus spp. stand volume in Zululand, South Africa: an analysis using a stochastic gradient boosting regression ensemble with multi-source data sets

Accurate, reliable, and up-to-date forest stand volume information is a prerequisite for a detailed evaluation of commercial forest resources and their sustainable management. Commercial forest responses to global climate change remain uncertain, and hence the mapping of stand volume as carbon sinks is fundamentally important in understanding the role of forests in stabilizing climate change effects. The aim of this study was to examine the utility of stochastic gradient boosting (SGB) and multi-source data to predict stand volume of a Eucalyptus plantation in South Africa. The SGB ensemble, random forest (RF), and stepwise multiple-linear regression (SMLR) were used to predict Eucalyptus stand volume and other related tree-structural attributes such as mean tree height and mean diameter at breast height (DBH). Multi-source data consisted of SPOT-5 raw spectral features (four bands), 14 spectral vegetation indices, rainfall data, and stand age. When all variables were used, the SGB algorithm showed that stand volume can be accurately estimated (R2 = 0.78 and RMSE = 33.16 m3 ha−1 (23.01% of the mean)). The competing RF ensemble produced an R2 value of 0.76 and a RMSE value of 37.28 m3 ha−1 (38.28% of the mean). SMLR on the other hand, produced an R2 value of 0.65 and an RMSE value of 42.50 m3 ha−1 (42.50% of the mean). Our study further showed that Eucalyptus mean tree height (R2 = 0.83 and RMSE = 1.63 m (9.08% of the mean)) and mean diameter at breast height (R2 = 0.74 and RMSE = 1.06 (7.89% of the mean)) can also be reasonably predicted using SGB and multi-source data. Furthermore, when the most important SGB model-selected variables were used for prediction, the predictive accuracies improved significantly for mean DBH (R2 = 0.81 and RMSE = 1.21 cm (6.12% of the mean)), mean tree height (R2 = 0.86 and RMSE = 1.39 m (7.02% of the mean)), and stand volume (R2 = 0.83 and RMSE = 29.58 m3 ha−1 (17.63% of the mean)). These results underscore the importance of integrating multi-source data with remotely sensed data for predicting Eucalyptus stand volume and related tree-structural attributes.

Assessing the potential of integrated Landsat 8 thermal bands, with the traditional reflective bands and derived vegetation indices in classifying urban landscapes

Abstract Reliable and up-to-date urban land cover information is valuable in urban planning and policy development. Due to the increasing demand for reliable land cover information there has been a growing need for robust methods and datasets to improve the classification accuracy from remotely sensed imagery. This study sought to assess the potential of the newly launched Landsat 8 sensor’s thermal bands and derived vegetation indices in improving land cover classification in a complex urban landscape using the support vector machine classifier. This study compared the individual and combined performance of Landsat 8’s reflective, thermal bands and vegetation indices in classifying urban land use-land cover. The integration of Landsat 8 reflective bands, derived vegetation indices and thermal bands overall produced significantly higher accuracy classification results than using traditional bands as standalone (i.e. overall, user and producer accuracies). An overall accuracy above 89.33% and a kappa index of 0.86, significantly higher than the one obtained with the use of the traditional reflective bands as a standalone data-set and other analysis stages. On average, the results also indicate high producer and user accuracies (i.e. above 80%) for most of the classes with a McNemar’s Z score of 9.00 at 95% confidence interval showing significant improvement compared with classification using reflective bands as standalone. Overall, the results of this study indicate that the integration of the Landsat 8’s OLI and TIR data presents an invaluable potential for accurate and robust land cover classification in a complex urban landscape, especially in areas where the availability of high resolution datasets remains a challenge.

Water quality monitoring in sub-Saharan African lakes: a review of remote sensing applications

Water quality deterioration in sub-Saharan Africa has attained a scale that requires scientific intervention. It is therefore important to devise appropriate and reliable techniques to investigate the water quality of lakes and reservoirs for the development of water resource management strategies. Whilst conventional water quality monitoring methods have been widely used due to their accuracy, these methods are time-consuming, costly and practically impossible to use at broader scales. This paper reviews the literature on various remote sensing platforms and techniques used for assessing and monitoring water quality in sub-Saharan Africa, and highlights their strengths and weaknesses. The use of remote sensing technology could enhance water quality monitoring, since remotely sensed data offer timely, up-to-date and comparatively accurate information, which is necessary for water resource management and strategic decision making. However, the use of this technology in some parts of sub-Saharan Africa is still at its infancy because of its high cost and limited availability.

Estimating wood volume from canopy area in deciduous woodlands of Zimbabwe

In this study we tested the predictive ability of canopy area in estimating wood volume in deciduous woodlands of Zimbabwe. The study was carried out in four sites of different climatic conditions. We used regression analysis to statistically quantify the prediction of wood volume from canopy area at species and stand level using field data. Our results revealed that canopy area significantly (P < 0.05) predicted wood volume at both levels. Furthermore, the results show that at the species-specific level, canopy area explained 54–81% of the variance in wood volume with standard error of estimate (SEE) ranging from 0.056 to 0.71 m3. At stand level, canopy area significantly (P < 0.05) explained 58–84% of the variance in total wood volume with SEE ranging from 0.15 to 3.99 m3 ha−1. Across all study sites, the relationship between canopy area and wood volume at stand level was best described by a logistic regression function, with a R2 value of 0.65 and SEE of 0.7 m3. We concluded that canopy area significantly (P < 0.05) predicted wood volume of dominant tree species in Zimbabwean deciduous woodlands. The relationship between wood volume and canopy area provides an opportunity of estimating wood volume using remote sensing as canopy area can be viewed and measured from aerial, as well as satellite-borne sensors.

Separability of coffee leaf rust infection levels with machine learning methods at Sentinel-2 MSI spectral resolutions

Coffee leaf rust (CLR) caused by the fungus Hemileia vastarix is a devastating disease in almost all coffee producing countries and remote sensing approaches have the potential to monitor the disease. This study evaluated the potential of Sentinel-2 band settings for discriminating CLR infection levels at leaf levels. Field spectra were resampled to the band settings of the Sentinel-2, and evaluated using the random forest (RF) and partial least squares discriminant analysis (PLS-DA) algorithms with and without variable optimization. Using all variables, Sentinel-2 Multispectral Imager (MSI)-derived vegetation indices achieved higher overall accuracy of 76.2% when compared to 69.8% obtained using raw spectral bands. Using the RF out-of-bag (OOB) scores, 4 spectral bands and 7 vegetation indices were identified as important variables in CLR discrimination. Using the PLS-DA Variable Importance in Projection (VIP) score, 3 Sentinel-2 spectral bands (B4, B6 and B5) and 5 vegetation indices were found to be important variables. Use of the identified variables improved the CLR discrimination accuracies to 79.4 and 82.5% for spectral bands and indices respectively when discriminated with the RF. Discrimination accuracy slightly increased through variable optimization for PLS-DA using spectral bands (63.5%) and vegetation indices (71.4%). Overall, this study showed the potential of the Sentinel 2 MSI band settings for CLR discrimination as part of crop condition assessment. Nevertheless further studies are required under field conditions.

An assessment of gully erosion along major armoured roads in south-eastern region of South Africa: a remote sensing and GIS approach

An assessment of gully erosion along road drainage-release sites is critical for understanding the contribution of roads to soil loss and for informed land management practices. Considering that road-related gully erosion has traditionally been measured using field methods that are expensive, tedious and limited spatially as well as temporally, it is important to identify affordable, timely and robust methods that can be used to effectively map and estimate the volume of gullies along the road networks. In this study, gullies along major roads were identified from remotely sensed data sets and their volumes were estimated in a Geographic Information Systems environment. Also, the biophysical and climatic factors such as vegetation cover, the road contributing surface area, the gradient of the discharge hillslope and rainfall were derived from remotely sensed data sets using Geographic Information Systems techniques to find out whether they could explain the morphology of gullies that existed in this area. The results of this study indicate that hillslope gradient (R2 = 0.69, α = 0.00) and road contributing surface area (R2 = 0.63, α = 0.00) have a strong influence on the volume of gullies along the major roads in the south-eastern region of South Africa, as might have been expected. However, other factors such as vegetation cover (R2 = 0.52, α = 0.00) and rainfall (R2 = 0.41 and α = 0.58) have a moderately weaker influence on the overall volume of gullies. Overall, the findings of this study highlight the importance of using remote sensing and Geographic Information Systems technologies in investigating gully erosion occurrence along major roads where detailed field work remains a challenge.

Applying the Surface Energy Balance System (SEBS) remote sensing model to estimate spatial variations in evapotranspiration in Southern Zimbabwe

Accurate, reliable and continuous understanding of water utilisation by different land cover types in arid environments is critical for water loss accounting to ensure sustainable water management in the face of the changing climate. Remote sensing provides a lucrative alternative for mapping and estimating the spatial and temporal distribution of water loss across the catchment. The results of this study have shown that evapotranspiration (ET) can be accurately estimated in arid environments from remotely sensed data, such as Moderate Resolution Imaging Spectroradiometer (MODIS) data, based on the Surface Energy Balance System (SEBS) algorithm. This study observed significant spatial and temporal variations in ET across the south western part of Zimbabwe. The findings from this study, therefore, underscore the importance of using cheap and readily available remotely sensed data for estimating and mapping the variations in ET in arid-environment areas found mainly in developing countries.