Abel Ramoelo

Regional estimation of savanna grass nitrogen using the red-edge band of the spaceborne RapidEye sensor

The regional mapping of grass nutrients is of interest in the sustainable planning and management of livestock and wildlife grazing. The objective of this study was to estimate and map foliar and canopy nitrogen (N) at a regional scale using a recent high resolution spaceborne multispectral sensor (i.e. RapidEye) in the Kruger National Park (KNP) and its surrounding areas, South Africa. The RapidEye sensor contains five spectral bands in the visible-to-near infrared (VNIR), including a red-edge band centered at 710 nm. The importance of the red-edge band for estimating foliar chlorophyll and N concentrations has been demonstrated in many previous studies, mostly using field spectroscopy. The utility of the red-edge band of the RapidEye sensor for estimating grass N was investigated in this study. A two-step approach was adopted involving (i) vegetation indices and (ii) the integration of vegetation indices with environmental or ancillary variables using a stepwise multiple linear regression (SMLR) and a non-linear spatial least squares regression (PLSR). The model involving the simple ratio (SR) index (R805/R710) defined as SR54, altitude and the interaction between SR54 and altitude (SR54 * altitude) yielded the highest accuracy for canopy N estimation, while the non-linear PLSR yielded the highest accuracy for foliar N estimation through the integration of remote sensing (SR54) and environmental variables. The study demonstrated the possibility to map grass nutrients at a regional scale provided there is a spaceborne sensor encompassing the red edge waveband with a high spatial resolution.

Validation of Global Evapotranspiration Product (MOD16) using Flux Tower Data in the African Savanna, South Africa

Globally, water is an important resource required for the survival of human beings. Water is a scarce resource in the semi-arid environments, including South Africa. In South Africa, several studies have quantified evapotranspiration (ET) in different ecosystems at a local scale. Accurate spatially explicit information on ET is rare in the country mainly due to lack of appropriate tools. In recent years, a remote sensing ET product from the MODerate Resolution Imaging Spectrometer (MOD16) has been developed. However, its accuracy is not known in South African ecosystems. The objective of this study was to validate the MOD16 ET product using data from two eddy covariance flux towers, namely; Skukuza and Malopeni installed in a savanna and woodland ecosystem within the Kruger National Park, South Africa. Eight day cumulative ET data from the flux towers was calculated to coincide with the eight day MOD16 products over a period of 10 years from 2000 to 2010. The Skukuza flux tower results showed inconsistent comparisons with MOD16 ET. The Malopeni site achieved a poorer comparison with MOD16 ET compared to the Skukuza, and due to a shorter measurement period, data validation was performed for 2009 only. The inconsistent comparison of MOD16 and flux tower-based ET can be attributed to, among other things, the parameterization of the Penman-Monteith model, flux tower measurement errors, and flux tower footprint vs.

Monitoring grass nutrients and biomass as indicators of rangeland quality and quantity using random forest modelling and WorldView-2 data

Land use and climate change could have huge impacts on food security and the health of various ecosystems. Leaf nitrogen (N) and above-ground biomass are some of the key factors limiting agricultural production and ecosystem functioning. Leaf N and biomass can be used as indicators of rangeland quality and quantity. Conventional methods for assessing these vegetation parameters at landscape scale level are time consuming and tedious. Remote sensing provides a bird-eye view of the landscape, which creates an opportunity to assess these vegetation parameters over wider rangeland areas. Estimation of leaf N has been successful during peak productivity or high biomass and limited studies estimated leaf N in dry season. The estimation of above-ground biomass has been hindered by the signal saturation problems using conventional vegetation indices. The objective of this study is to monitor leaf N and above-ground biomass as an indicator of rangeland quality and quantity using WorldView-2 satellite images and random forest technique in the north-eastern part of South Africa. Series of field work to collect samples for leaf N and biomass were undertaken in March 2013, April or May 2012 (end of wet season) and July 2012 (dry season). Several conventional and red edge based vegetation indices were computed. Overall results indicate that random forest and vegetation indices explained over 89% of leaf N concentrations for grass and trees, and less than 89% for all the years of assessment. The red edge based vegetation indices were among the important variables for predicting leaf N. For the biomass, random forest model explained over 84% of biomass variation in all years, and visible bands including red edge based vegetation indices were found to be important. The study demonstrated that leaf N could be monitored using high spatial resolution with the red edge band capability, and is important for rangeland assessment and monitoring.

Potential of Sentinel-2 spectral configuration to assess rangeland quality

Abstract. Sentinel-2 is intended to improve vegetation assessment at local to global scales. Today, estimation of leaf nitrogen (N) as an indicator of rangeland quality is possible using hyperspectral systems. However, few studies based on commercial imageries have shown a potential of the red-edge band to accurately predict leaf N at the broad landscape scale. We intend to investigate the utility of Sentinel-2 for estimating leaf N concentration in the African savanna. Grass canopy reflectance was measured using the analytical spectral device (ASD) in concert with leaf sample collections for leaf N chemical analysis. ASD reflectance data were resampled to the spectral bands of Sentinel-2 using published spectral response functions. Random forest (RF), partial least square regression (PLSR), and stepwise multiple linear regression (SMLR) were used to predict leaf N using all 13 bands. Using leave-one-out cross validation, the RF model explained 90% of leaf N variation, with a root mean square error of 0.04 (6% of the mean), which is higher than that of PLSR and SMLR. Using RF, spectral bands centered at 705 nm (red edge) and two shortwave infrared bands centered at 2190 and 1610 nm were found to be the most important bands in predicting leaf N.

Savanna grass nitrogen to phosphorous ratio estimation using field spectroscopy and the potential for estimation with imaging spectroscopy

Determining the foliar N:P ratio provides a tool for understanding nutrient limitation on plant production and consequently for the feeding patterns of herbivores. In order to understand the nutrient limitation at landscape scale, remote sensing techniques offer that opportunity. The objective of this study is to investigate the utility of field spectroscopy and a potential of hyperspectral mapper (HyMap) spectra to estimate foliar N:P ratio. Field spectral measurements were undertaken, and grass samples were collected for foliar N and P extraction. The foliar N:P ratio prediction models were developed using partial least square regression (PLSR) with original spectra and transformed spectra for field and the resampled field spectra to HyMap. Spectral transformations included the continuum removal (CR), water removal (WR), first difference derivative (FD) and log transformation (Log(1/R)). The results showed that CR and WR spectra in combination with PLSR predicted foliar N:P ratio with higher accuracy as compared to FD and R, using field spectra. For HyMap spectral analysis, addition to CR and WR, FD achieved higher estimation accuracy. The performance of FD, CR and WR spectra were attributed to their ability to minimize sensor and water effects on the fresh leaf spectra, respectively. The study demonstrated a potential to predict foliar N:P ratio using field and HyMap simulated spectra and shortwave infrared (SWIR) found to be highly sensitive to foliar N:P ratio. The study recommends the prediction of foliar N:P ratio at landscape level using airborne hyperspectral data and could be used by the resource managers, park managers, farmers and ecologists to understand the feeding patterns, resource selection and distribution of herbivores (i.e. wild and livestock).

Seasonal separation of African savanna components using WorldView-2 imagery : a comparison of pixel- and object-based approaches and selected classification algorithms

Separation of savanna land cover components is challenging due to the high heterogeneity of this landscape and spectral similarity of compositionally different vegetation types. In this study, we tested the usability of very high spatial and spectral resolution WorldView-2 (WV-2) imagery to classify land cover components of African savanna in wet and dry season. We compared the performance of Object-Based Image Analysis (OBIA) and pixel-based approach with several algorithms: k-nearest neighbor (k-NN), maximum likelihood (ML), random forests (RF), classification and regression trees (CART) and support vector machines (SVM). Results showed that classifications of WV-2 imagery produce high accuracy results (>77%) regardless of the applied classification approach. However, OBIA had a significantly higher accuracy for almost every classifier with the highest overall accuracy score of 93%. Amongst tested classifiers, SVM and RF provided highest accuracies. Overall classifications of the wet season image provided better results with 93% for RF. However, considering woody leaf-off conditions, the dry season classification also performed well with overall accuracy of 83% (SVM) and high producer accuracy for the tree cover (91%). Our findings demonstrate the potential of imagery like WorldView-2 with OBIA and advanced supervised machine-learning algorithms in seasonal fine-scale land cover classification of African savanna.

Spectral variability within species and its effects on Savanna tree species discrimination

Differences in within-species phenology and structure driven by factors including topography, edaphic properties, and climatic variables present important challenges for species differentiation with remote sensing in the Kruger National Park, South Africa. The objective of this study was to examine probable factors including intraspecies spectral variability and the spectral sample size that could affect remote sensing of Savanna tree species across a land-use gradient in the Kruger National park. Eighteen species were examined: Acacia gerradii, Acacia nigrescens, Combretum apiculatum, Combretum collinum, Combretum hereroense, Combretum imberbe, Combretum zeyheri, Dichrostachys cinerea, Euclea sp (E. divinurum and E. natalensis, Gymnosporia sp (G. buxifolia and G. senegalensis), Lonchocarpus capassa, Peltoforum africanum, Piliostigma thonningii, Pterocarpus rotundifolia, Sclerocarya birrea, Strychnos sp (S. madagascariensis, S. usambarensis), Terminalia sericea and Ziziphus mucronata. Discriminating species using the K-nearest neighbour (K = 1) classifier with spectral angle mapper (SAM) yielded a higher classification accuracy (48% overall accuracy) compared to 16% for the classification involving the mean spectra for each species as the training spectral set. Within-species spectral variability and the training sample size were identified as important factors affecting classification accuracy of the tree species. We recommend a non-parametric classifier such as K-nearest neighbour classifier for classifying and mapping tree species in a highly complex environment such as the savanna system of the Kruger National Park.

Comparison of Landsat 8 OLI and Landsat 7 ETM+ for estimating grassland LAI using model inversion and spectral indices: case study of Mpumalanga, South Africa

ABSTRACT The leaf area index (LAI) is the key biophysical indicator used to assess the condition of rangeland. In this study, we investigated the implications of narrow spectral response, high radiometric resolution (12 bits), and higher signal-to-noise ratio of the Landsat 8 Operational Land Imager (OLI) sensor for the estimation of LAI. The Landsat 8 LAI estimates were compared to that of its predecessors, namely Landsat 7 Enhanced Thematic Mapper Plus (ETM+) (8 bits). Furthermore, we compared the radiative transfer model (RTM) and spectral indices approaches for estimating LAI on rangeland systems in South Africa. The RTM was inverted using artificial neural network (ANN) and lookup table (LUT) algorithms. The accuracy of the models was higher for Landsat 8 OLI, where ANN (root mean squared error, RMSE = 0. 13; R2 = 0. 89), LUT (RMSE = 0. 25; R2 = 0. 50), compared to Landsat 7 ETM+, where ANN (RMSE = 0. 35; R2 = 0. 60), LUT (RMSE = 0. 38; R2 = 0. 50). Compared to an empirical approach, the RTM provided higher accuracy. In conclusion, Landsat 8 OLI provides an improvement for the estimation of LAI over Landsat 7 ETM+. This is useful for rangeland monitoring.

An intercomparison of satellite-based daily evapotranspiration estimates under different eco-climatic regions in South Africa

Knowledge of evapotranspiration (ET) is essential for enhancing our understanding of the hydrological cycle, as well as for managing water resources, particularly in semi-arid regions. Remote sensing offers a comprehensive means of monitoring this phenomenon at different spatial and temporal intervals. Currently, several satellite methods exist and are used to assess ET at various spatial and temporal resolutions with various degrees of accuracy and precision. This research investigated the performance of three satellite-based ET algorithms and two global products, namely land surface temperature/vegetation index (TsVI), Penman–Monteith (PM), and the Meteosat Second Generation ET (MET) and the Global Land-surface Evaporation: the Amsterdam Methodology (GLEAM) global products, in two eco-regions of South Africa. Daily ET derived from the eddy covariance system from Skukuza, a sub-tropical, savanna biome, and large aperture boundary layer scintillometer system in Elandsberg, a Mediterranean, fynbos biome, during the dry and wet seasons, were used to evaluate the models. Low coefficients of determination (R2) of between 0 and 0.45 were recorded on both sites, during both seasons. Although PM performed best during periods of high ET at both sites, results show it was outperformed by other models during low ET times. TsVI and MET were similarly accurate in the dry season in Skukuza, as GLEAM was the most accurate in Elandsberg during the wet season. The conclusion is that none of the models performed well, as shown by low R2 and high errors in all the models. In essence, our results conclude that further investigation of the PM model is possible to improve its estimation of low ET measurements.

Estimation of leaf area index (LAI) of South Africa from MODIS imager by inversion of PROSAIL radiative transfer model

Over-grazing, bush encroachment and alien species invasion are having negative impacts on livestock production in the rangelands of Sub-Saharan Africa, thus threatening rural livelihoods in the region. Leaf area index (LAI) can be used as an indicator or an early warning signal of changes in rangeland conditions. The aims of this study were to (i) assess the accuracy of the existing MODIS LAI product for key land cover classes and biomes in South Africa and (ii) assess the accuracy of LAI retrieved by the inversion of PROSAIL radiative transfer model. The MODIS LAI estimates were particularly poor for grassland and Karoo biomes for the period under investigation (November 2012 to February 2013). More accurate estimates of LAI values (RMSE = 0.9) across three major biomes were generated by inverting model outputs from PROSAIL radiative transfer model on MODIS imagery.