Elfatih M. Abdel-Rahman

Random forest regression and spectral band selection for estimating sugarcane leaf nitrogen concentration using EO-1 Hyperion hyperspectral data

Nitrogen (N) is one of the most important limiting nutrients for sugarcane production. Conventionally, sugarcane N concentration is examined using direct methods such as collecting leaf samples from the field followed by analytical assays in the laboratory. These methods do not offer real-time, quick, and non-destructive strategies for estimating sugarcane N concentration. Methods that take advantage of remote sensing, particularly hyperspectral data, can present reliable techniques for predicting sugarcane leaf N concentration. Hyperspectral data are extremely large and of high dimensionality. Many hyperspectral features are redundant due to the strong correlation between wavebands that are adjacent. Hence, the analysis of hyperspectral data is complex and needs to be simplified by selecting the most relevant spectral features. The aim of this study was to explore the potential of a random forest (RF) regression algorithm for selecting spectral features in hyperspectral data necessary for predicting sugarcane leaf N concentration. To achieve this, two Hyperion images were captured from fields of 6–7 month-old sugarcane, variety N19. The machine-learning RF algorithm was used as a feature-selection and regression method to analyse the spectral data. Stepwise multiple linear (SML) regression was also examined to predict the concentration of sugarcane leaf N after the reduction of the redundancy in hyperspectral data. The results showed that sugarcane leaf N concentration can be predicted using both non-linear RF regression (coefficient of determination, R 2 = 0.67; root mean square error of validation (RMSEV) = 0.15%; 8.44% of the mean) and SML regression models (R 2 = 0.71; RMSEV = 0.19%; 10.39% of the mean) derived from the first-order derivative of reflectance. It was concluded that the RF regression algorithm has potential for predicting sugarcane leaf N concentration using hyperspectral data.

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.

The Utility of AISA Eagle Hyperspectral Data and Random Forest Classifier for Flower Mapping

Knowledge of the floral cycle and the spatial distribution and abundance of flowering plants is important for bee health studies to understand the relationship between landscape and bee hive productivity and honey flow. The key objective of this study was to show how AISA Eagle hyperspectral data and random forest (RF) can be optimally utilized to produce flowering and spatially explicit land use/land cover (LULC) maps for a study site in Kenya. AISA Eagle imagery was captured at the early flowering period (January 2014) and at the peak flowering season (February 2013). Data on white and yellow flowering trees as well as LULC classes in the study area were collected and used as ground-truth points. We utilized all 64 AISA Eagle bands and also used variable importance in RF to identify the most important bands in both AISA Eagle data sets. The results showed that flowering was most accurately mapped using the AISA Eagle data from the peak flowering period (85.71%–88.15% overall accuracy for the peak flowering season imagery versus 80.82%–83.67% for the early flowering season). The variable optimization (i.e., variable selection) analysis showed that less than half of the AISA bands (n = 26 for the February 2013 data and n = 21 for the January 2014 data) were important to attain relatively reliable classification accuracies. Our study is an important first step towards the development of operational flower mapping routines and for understanding the relationship between flowering and bees’ foraging behavior.

Estimating Swiss chard foliar macro- and micronutrient concentrations under different irrigation water sources using ground-based hyperspectral data and four partial least squares (PLS)-based (PLS1, PLS2, SPLS1 and SPLS2) regression algorithms

Swiss chard nutrients were estimated under trial conditions using hyperspectral data.Partial least squares (PLS1 and PLSR2) and sparse PLS1 and PLS2 regressions performance was compared.SPLS-based regression methods significantly reduced the dimensionality in hyperspectral data.Macronutrients estimation models were more accurate than micronutrients estimation ones. Timely information on crop foliar nutrient content provides a measure of crop nutritional and vitality status. Growers and farm managers use such information for precision crop management such as an appropriate fertilizer application to correct for any crop nutrient deficiencies at identified hotspots. Foliar heavy nutrient content could also be a direct indicator of crops having been polluted from the surroundings, which may be a result of heavy metals absorbed from, among others, contaminated soils and waste water. In the present study, we explored the potential use of four partial least squares (PLS)-based regression algorithms for estimating foliar Swiss chard macro- and micronutrient concentrations using ground-based hyperspectral data under three treatments; i.e. rainwater+fertilizer (R+F), tap water+fertilizer (T+F), and treated wastewater (W). Swiss chard canopy-level hyperspectral measurements under these three treatments were collected using a handheld spectroradiometer 2.5months after sowing. The reflectance spectra were normalized to their first-order derivatives. The concentrations of three Swiss chard foliar macronutrients (NPK) and three micronutrients (Zn, Cu and Fe) under the three treatments were determined. Regression models for estimating macro- and micronutrient concentrations were then derived using PLS1 and sparse PLS1 methods, while the potential simultaneous estimation of the macronutrient as well as micronutrient concentrations was explored using the PLS2 and SPLS2 regression approaches. Results showed that high variances in the macro- and micronutrient concentrations can be explained by the four regression models under the three treatments (R2train ranged between 0.73 and 0.99), except when P, Zn and Cu concentrations were estimated using the PLS2-based models under the three treatments (R2train ranged between 0.08 and 0.68) and Fe concentration using SPLSR1 under W treatment (R2train=0.64). Our results further showed that Swiss chard foliar N (RMSE=1.67%) concentration under R+F treatment and Fe (RMSE=7.83%) concentration under the T+F treatment most accurately estimated macro- and micronutrients. Our study also showed that the Swiss chard foliar macronutrient concentrations were more accurately estimated compared to micronutrient concentrations and PLS2 outperformed the PLS1 based regression model. The results of the current study pave the way for developing an effective foliar nutrient estimation routine suitable for monitoring Swiss chard nutrient status under different treatments.

Maize Cropping Systems Mapping Using RapidEye Observations in Agro-Ecological Landscapes in Kenya

Cropping systems information on explicit scales is an important but rarely available variable in many crops modeling routines and of utmost importance for understanding pests and disease propagation mechanisms in agro-ecological landscapes. In this study, high spatial and temporal resolution RapidEye bio-temporal data were utilized within a novel 2-step hierarchical random forest (RF) classification approach to map areas of mono- and mixed maize cropping systems. A small-scale maize farming site in Machakos County, Kenya was used as a study site. Within the study site, field data was collected during the satellite acquisition period on general land use/land cover (LULC) and the two cropping systems. Firstly, non-cropland areas were masked out from other land use/land cover using the LULC mapping result. Subsequently an optimized RF model was applied to the cropland layer to map the two cropping systems (2nd classification step). An overall accuracy of 93% was attained for the LULC classification, while the class accuracies (PA: producer’s accuracy and UA: user’s accuracy) for the two cropping systems were consistently above 85%. We concluded that explicit mapping of different cropping systems is feasible in complex and highly fragmented agro-ecological landscapes if high resolution and multi-temporal satellite data such as 5 m RapidEye data is employed. Further research is needed on the feasibility of using freely available 10–20 m Sentinel-2 data for wide-area assessment of cropping systems as an important variable in numerous crop productivity models.

Land Use and Land Cover Changes in Northern Kordofan State of Sudan: A Remotely Sensed Data Analysis

The North Kordofan region is semiarid and characterized by recurrent episodes of drought which led to increasing desertification. The agricultural and forest production in North Kordofan State (NKS), however, is adversely hampered by climate change, particularly the unreliable and fluctuated rainfall and desertification. Hence, it is expected that the land use/land cover (LULC) classes in the state would have dramatically changed during past decades. This study tries to detect the changes in LULC in NKS during the period between 1973 and 2001. We assess the desertification process using vegetation cover as an indicator. We used remotely sensed data from Landsat multispectral scanner (MSS; captured in 1973) and enhanced thematic mapper plus (ETM+; captured in 2001) to detect LULC conversion dynamics. Pre- and postclassification change detection methods were compared. A supervised image classification (maximum likelihood) is then performed to identify LULC classes. Ten major land cover classes are discriminated. These are forests, farms on sand, farms on clay, fallows on sand, fallows on clay, woodlands, mixed woodlands, grasslands, burnt/wetlands, and natural water bodies. The results revealed that using a preclassification image differencing procedure, positive (9.66 and 6.70 % of total area when near-infrared (NIR) and normalized difference vegetation index (NDVI) were used, respectively), negative (9.77 and 6.62 % of total area when NIR and NDVI were used, respectively), and no (80.57 and 86.68 % of total area when NIR and NDVI were used, respectively) vegetation changes were observed in the study area during the period 1973–2001. The study also indicates a negative change trend when principal component analysis (PCA) and change vector analysis (CVA) methods are employed. With respect to the postclassification method, the results show significant conversions in LULC classes, where new classes such as farms and fallows on clay soils were introduced in 2001, while woodlands in 1973 were completely shifted to farm on sand, farm on clay, fallow on sand, fallow on clay, grassland, and mixed woodland in 2001. The study demonstrates different signs of desertification in the study area related to change patterns in LULC classes, such as increase in farms on sand and clay soils at the expense of wood and grasslands. It is concluded that the vegetation cover in North Kordofan was negatively changed due to socioeconomic factors and desertification in the area was the main sign of such negative LULC changes.

Wide-area mapping of invasive species propagation and containment zones in somaliland using phenometric trends and generalized linear modelling

Spatial information on the occurrence and propagation of invasive species is imperative in order to manage their risk and spread. In this contribution we used phenology and vegetation productivity trends (2001 to 2014) from 250-meter MODIS (Moderate Resolution Imaging Spectroradiometer) EVI (Enhanced Vegetation Index) time-series data to map propagation and possible containment areas for Prosopis juliflora and Parthenium hysterophorus in western Somaliland (eastern Africa). Generalized Linear Modeling (GLM) with a binomial logistics function was used to link available reference data (Landsat-based) on both invasive species to the MODIS-based vegetation trends. Spread corridors and containment zones were, furthermore, identified for both species. Variable relevance in GLM showed that the variables ‘EVI trend’ and ‘peak value’ were highly relevant for P juliflora (log odds ratio >200, p<0.001 and regression estimate >|5|). Riverside and peri-urban areas were identified as important propagation and risk zones.

Potential distribution dataset of honeybees in Indian Ocean Islands: Case study of Zanzibar Island

Honeybees (Apis mellifera) are principal insect pollinators, whose worldwide distribution and abundance is known to largely depend on climatic conditions. However, the presence records dataset on potential distribution of honeybees in Indian Ocean Islands remain less documented. Presence records in shape format and probability of occurrence of honeybees with different temperature change scenarios is provided in this article across Zanzibar Island. Maximum entropy (Maxent) package was used to analyse the potential distribution of honeybees. The dataset provides information on the current and future distribution of the honey bees in Zanzibar Island. The dataset is of great importance for improving stakeholders understanding of the role of temperature change on the spatial distribution of honeybees.