Probabilistic Mapping and Spatial Pattern Analysis of Grazing Lawns in Southern African Savannahs Using WorldView-3 Imagery and Machine Learning Techniques

• Main Authors: Kwame T. Awuah, Paul Aplin, Christopher G. Marston, Ian Powell, Izak P. J. Smit
• Format: Article
• Language: English
• Published: MDPI AG 2020-10-01
• Series: Remote Sensing
• Subjects: African savannah; grazing lawns; machine learning; WorldView-3; Support Vector Machines; Random Forest
• Online Access: https://www.mdpi.com/2072-4292/12/20/3357

Savannah grazing lawns are a key food resource for large herbivores such as blue wildebeest (<i>Connochaetes taurinus</i>), hippopotamus (<i>Hippopotamus amphibius</i>) and white rhino (<i>Ceratotherium simum</i>), and impact herbivore densities, movement and recruitment rates. They also exert a strong influence on fire behaviour including frequency, intensity and spread. Thus, variation in grazing lawn cover can have a profound impact on broader savannah ecosystem dynamics. However, knowledge of their present cover and distribution is limited. Importantly, we lack a robust, broad-scale approach for detecting and monitoring grazing lawns, which is critical to enhancing understanding of the ecology of these vital grassland systems. We selected two sites in the Lower Sabie and Satara regions of Kruger National Park, South Africa with mesic and semiarid conditions, respectively. Using spectral and texture features derived from WorldView-3 imagery, we (i) parameterised and assessed the quality of Random Forest (RF), Support Vector Machines (SVM), Classification and Regression Trees (CART) and Multilayer Perceptron (MLP) models for general discrimination of plant functional types (PFTs) within a sub-area of the Lower Sabie landscape, and (ii) compared model performance for probabilistic mapping of grazing lawns in the broader Lower Sabie and Satara landscapes. Further, we used spatial metrics to analyse spatial patterns in grazing lawn distribution in both landscapes along a gradient of distance from waterbodies. All machine learning models achieved high F-scores (F1) and overall accuracy (OA) scores in general savannah PFTs classification, with RF (F1 = <inline-formula><math display="inline"><semantics><mrow><mn>95.73</mn><mo>±</mo><mn>0.004</mn><mo>%</mo></mrow></semantics></math></inline-formula>, OA = <inline-formula><math display="inline"><semantics><mrow><mn>94.16</mn><mo>±</mo><mn>0.004</mn><mo>%</mo></mrow></semantics></math></inline-formula>), SVM (F1 = <inline-formula><math display="inline"><semantics><mrow><mn>95.64</mn><mo>±</mo><mn>0.002</mn><mo>%</mo></mrow></semantics></math></inline-formula>, OA = <inline-formula><math display="inline"><semantics><mrow><mn>94.02</mn><mo>±</mo><mn>0.002</mn><mo>%</mo></mrow></semantics></math></inline-formula>) and MLP (F1 = <inline-formula><math display="inline"><semantics><mrow><mn>95.71</mn><mo>±</mo><mn>0.003</mn><mo>%</mo></mrow></semantics></math></inline-formula>, OA = <inline-formula><math display="inline"><semantics><mrow><mn>94.27</mn><mo>±</mo><mn>0.003</mn><mo>%</mo></mrow></semantics></math></inline-formula>) forming a cluster of the better performing models and marginally outperforming CART (F1 = <inline-formula><math display="inline"><semantics><mrow><mn>92.74</mn><mo>±</mo><mn>0.006</mn><mo>%</mo></mrow></semantics></math></inline-formula>, OA = <inline-formula><math display="inline"><semantics><mrow><mn>90.93</mn><mo>±</mo><mn>0.003</mn><mo>%</mo></mrow></semantics></math></inline-formula>). Grazing lawn detection accuracy followed a similar trend within the Lower Sabie landscape, with RF, SVM, MLP and CART achieving F-scores of 0.89, 0.93, 0.94 and 0.81, respectively. Transferring models to the Satara landscape however resulted in relatively lower but high grazing lawn detection accuracies across models (RF = 0.87, SVM = 0.88, MLP = 0.85 and CART = 0.75). Results from spatial pattern analysis revealed a relatively higher proportion of grazing lawn cover under semiarid savannah conditions (Satara) compared to the mesic savannah landscape (Lower Sabie). Additionally, the results show strong negative correlation between grazing lawn spatial structure (fractional cover, patch size and connectivity) and distance from waterbodies, with larger and contiguous grazing lawn patches occurring in close proximity to waterbodies in both landscapes. The proposed machine learning approach provides a novel and robust workflow for accurate and consistent landscape-scale monitoring of grazing lawns, while our findings and research outputs provide timely information critical for understanding habitat heterogeneity in southern African savannahs.