it_team4_FOREst 28 March 2021HNEE FIT graduate André Parplies analysed during his research period phenometric of semi-arid grasslands in South Africa.
He studied spatial patterns of South African rangelands on farm scale using a phenometrical approach and investigated the following research question:
- How is the variability within and between different farm units?
- Is it possible to discriminate between different tenure systems (i.e. commercial vs. communal) and which phenological parameters are most beneficial?
He used high resolution RapidEye imagery covering two vegetation growth seasons from 2011 till 2013 to derive key phenometrics for a study area located in the semi-arid grassland biome in South Africa. A state of the art pre-processing technique for the standardization of remotely sensed satellite images was implemented as a basis for further processing, like noise reduction and the creation of a Normalized Difference Vegetation Index (NDVI) time series.
Quantitative phenology parameters were derived and visualized showing the start, end, length, amplitude and the small integral as a proxy for vegetation productivity of the two growing seasons. Each season was detected successfully with starting and ending points around October/November and around July and August respectively.
Descriptive statistics as well as the visual interpretation of produced maps on regional and farm scale showed patterns allowed a separation between the two tenure systems. Variances could be visualized by ordinary kriging that showed individual situation of each farm camp. Most prominent phenometrics to point out differences between communally and commercially managed rangelands were the length, amplitude and small integral of the growing season (only amplitude shown here). The results proved to be useful by detecting spatial patterns and variances on small scale at farm level in the study area. The applied phenometrical approach generated promising and reasonable results by visualizing the spatial composition of different land use and land cover types in different tenure systems (commercial vs. communal farms). Recommendations and limitations for similar studies dealing with time series to derive seasonal information regarding the temporal resolution and data validation could be made.
The poster was submitted at AARSE 10TH CONFERENCE, AARSE2014 in Johannesburg and presented at ISRSE in Berlin 2015.
André Parplies published further articles on semi-arid grassland.
This work was conducted in cooperation of the working groups at Institute of Crop Science and Resource Conservation (INRES) and Centre for Remote Sensing of Land Surfaces (ZFL) in Bonn and the University for Sustainable Development (HNEE) in Eberswalde, Germany.