Premium
Method for discriminating objectively between patches and non‐patches in a semiarid environment
Author(s) -
Du Toit Justin C.O.,
Danckwerts Jock E.,
Zacharias Peter J.K.
Publication year - 2007
Publication title -
grassland science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.388
H-Index - 19
eISSN - 1744-697X
pISSN - 1744-6961
DOI - 10.1111/j.1744-697x.2007.00074.x
Subject(s) - ordination , range (aeronautics) , vegetation (pathology) , grazing , pasture , ecology , scale (ratio) , canonical correspondence analysis , environmental science , geography , biology , species richness , cartography , medicine , materials science , pathology , composite material
Grass swards that have been grazed by livestock exhibit a patchy structure comprising short and tall components. While such patches are often observed, they are seldom objectively described. This study sought to develop a technique for objectively discriminating between patches (short) and non‐patches (tall) in a semiarid savanna in the Eastern Cape, South Africa. Two independent hypotheses were addressed: (i) vegetation can be categorized into patches and non‐patches by analyzing data comprising a large number of single measurements of sward height; and (ii) there are two distinct grass communities present, each of which comprises a particular range of sward heights. Single and double normal distributions were fitted to height distributions using the maximum likelihood estimation (MLE; hypothesis 1), and species composition data were ordinated using canonical correspondence analysis (CCA) and two‐way indicator species analysis (TWINSPAN; hypothesis 2). In general, the results supported both hypotheses. Patches and non‐patches can be distinguished using MLE unless grazing is patchy at a scale smaller than the lowest scale of sward measurement (here 0.5 m, the diameter of a disc pasture meter). Ordination and classification revealed that – at our site – two distinct grass communities are present, and that patches and non‐patches are characterized by vegetation <6 and 6 cm, respectively.