The relationship between multispectral satellite-derived vegetation indices and forage quantity and quality indicators in Mountain Zebra National Park

Thesis (MScAgric)--Stellenbosch University, 2019.

ENGLISH ABSTRACT: Climate change and human population growth has put pressure on protected areas and wilderness in southern Africa, consequently limiting food availability for herbivores. This is a major concern as forage quantity and quality available to herbivores affect the health and dynamics of herbivore populations. Therefore, the monitoring and assessment of forage quantity and quality across a landscape can potentially help ecologists make appropriate management decisions for herbivore populations in protected areas. Remote sensing measuring techniques, specifically multispectral satellite-derived vegetation indices (VI’s), can be a useful tool in providing information and knowledge about forage quantity and quality for herbivores, and the possible changes in these resources. However, for the accurate interpretation and implementation of multispectral satellite-derived VI’s, their relationship between forage quantity and quality indicators within the desired area needs to be determined. Therefore, this study assessed the relationship between multispectral remote sensors, onboard the MODIS and Sentinel-2 satellites, and forage quantity and quality indicators in MZNP. Forage biomass was estimated and used as a forage quantity indicator and forage nutrients were analysed and used as an indicator for forage quality. Correlations and regression techniques were applied, and results showed different multispectral satellite-derived VI’s have, at different strengths, relationships with the different forage quantity and quality indicators. The MODIS Normalized Difference Vegetation Index (NDVI), rather than the MODIS Enhanced Vegetation Index (EVI), showed a strong relationship with biomass and was more related to the amount of high forage quality in MZNP. The Sentinel-2 Chlorophyll Red-Edge index (Chlred-edge) had a very strong relationship with forage quality indicator total N concentration. Strong relationships were also found between the Green Chlorophyll Index (Clgreen), NDVI and MODIS NDVI and the forage quality indicators fiber (NDF, NDFd and ADL) and potassium. Herbivore faecal samples from the dominant ungulate species in MZNP were also analysed for forage quality indicators and related to the multispectral satellite-derived VI’s, to determine the relationship between VI’s and the diets of the herbivores in MZNP. Results showed associations between the MODIS NDVI and EVI can be generally related to dietary nitrogen, phosphorus and magnesium for the park’s dominant ungulate species. The research in this study also found valuable information on the relationship between remote sensing and forage quantity and quality, which up to now has lacked sufficient research. This study shows the implementation of multispectral satellite-derived VI’s can assist with the monitoring and assessment of forage quantity and quality for the herbivores of MZNP and will aid in making appropriate herbivore management decisions. The information uncovered by this study also demonstrates relationships between VI’s and forage quantity and quality that require improved understanding across a wider range of ecosystems.

AFRIKAANSE OPSOMMING: Klimaatsverandering en bevolkingsaanwas plaas al grooter druk op bewaringsgebiede in Suidelike-Afrika wat ‘n bedreiging ten opsigte van die beskikbaarheid van weiding vir weidende diere kan lei. Hierdie is ‘n bron van kommer aangesien die kwantiteit en kwaliteit van weiding onontbeerlik vir die instandhouding van diereprestasie en -gesondheid van weidende diere is. Daarom kan die monitering en bepaling van ruvoerkwaliteit en -kwantiteit van veld moontlik bydrae om ekoloë te help om bestuursbesluite ten opsigte van weidende diere in beskemde gebiede te maak. Afstands-sensor-metingstegnologie, spesifiek “multispectral satellite-derived vegetation indices (VI’s)”, is nuttige tegnologie om inligting en kennis oor beskikbare ruvoerkwaliteit en -kwantiteit vir herkouers asook die moontlike variasie daarvan te verskaf. ‘n Verhouding tussen weidingskwaliteit en -kwantiteit indikator binne die verlangde gebied moet bepaal word ten einde akurate interpretasie en implementering van “multispectral satellite-derived VI’s” moontlik te maak. Hierdie studie het dus die verhouding tussen “multispectral” afstand sensors aan boord die MODIS en Sentinal-2 sateliette en die weidingskwaliteit en -kwantiteit indikatore in MZNP beoordeel. Die weidingsbiomassa is geskat en gebruik as aanduiding van weidingskwaliteit terwyl die weidingsnutrïentanalises as aanduiding van die kwaliteit gedien het. Korrelasie en regressietegnieke is toegepas waarna resultate verskille in weidingskwaliteit en -kwantiiteit indikatore deur verskille in strekte en verhoudings van “multispectral satellite-derived VI’s.” uitgewys het. Die “MODIS Normalized Difference Vegetation Index (NDVI)” was meer akkuraat om die hoeveelheid hoë kwaliteit weiding in MZNP uit te wys en het ‘n hoër verwantskap teenoor die MODIS Enhanced Vegetation Index (EVI) getoon. Die Sentinel-2 Chlorophyll Red-Edge index (Chlred-edge) het ‘n sterk verwantskap met ruvoerkwaliteit, veral totale N-konsentrasie, getoon. Verdere sterk verwantskappe is tussen Green Chlorophyll Index (Clgreen), NDVI en MODIS NDVI en ruvoer-eienskappe (NBV, NBVd en SBV) en kalium bepaal. Mismonsters van die vernaamste wildspesies in MZNP is ook vir ruvoerkwaliteit ontleed en met die multispectral satellite-derived VI inligting vergelyk om soedoende die verwantskap tussen VI’s sowel as die diet van herkouers te ondersoek. ‘n Positiewe verwantskap tussen MODIS NDVI and EVI is bepaal en hou in die algemeen verband met dieetstikstof, -fosfor en -magnesium. Navorsing van hierdie studie dra waardevolle inligting by tot die verwantskap tussen afstandsmeting en ruvoerkwaliteit en -kwantiteit wat tot op hede ontoereikend was. Hierdie studie het aangetoon dat multispectral satellite-derived VI’s” kan bydrae tot die monitering en beoordeling van ruvoerkwaliteit en -kwantiteit van herkouers in MZNP. Hierdie tegnologie kan dus help om korrekte wildsbestuursbesluite te neem. Inligting van hierdie studie demonsteer verder ‘n verwantskap tussen VI’s en ruvoerkwaliteit en -kwantiteit in ag genome ‘n verbeterde begrip van die wyer ekosisteem.

Masters