Using GIS techniques to model the distribution of the economically important xate palm Chamaedorea ernesti-augusti within the Greater Maya Mountains, Belize

作者：Penn Malcolm G^*; Moncrieff Clive B; Bridgewater Samuel G M; Garwood Nancy C; Bateman Richard M; Chan Ivis; Cho Percival

来源：Systematics and Biodiversity, 2009, 7(1): 63-72.

DOI：10.1017/S1477200008002880

• Abundance
• Chamaedorea
• geostatistics
• interpolation
• kriging
• kriging range
• non-timber forest product
• spatial-variation
• variogram
• xate

The purpose of this investigation was to model the distribution of the fishtail xate palm (Chamaedorea ernesti-augusti) and to investigate the effectiveness of using geostatistical techniques to map its distribution. The leaves of the Central American xate palm C. ernesti-augusti are now extensively cut for the flori-cultural industry; across its range, populations are being over-harvested. In Central America, the Greater Maya Mountains (GMM), Belize are widely believed to be an abundance 'hotspot' for this species. In total, 220 sample sites were systematically located throughout the GMM, based on a fixed distance interval using the accessible track/road network. Species counts and environmental data were collected from each 20 M X 20 M plot. Statistical analyses were carried out using GenStat and Arc-GIS. We used several semivariogram models, within bespoke GenStat procedures, to map the log-transformed xate abundance data. Validation of models used the official GenStat Kcross-validation procedure. Of the six models investigated in detail, the 'Bounded Linear' model outperformed all others and most effectively represented the spatial distribution of C ernesti-augusti. Environmental factors were found to have low statistical significance for the distribution of the xate palm. The predicted map shows that the geographic abundance was dominated by low xate counts; xate abundance values for this region appear to have been consistently overestimated in the literature. Our study showed the effectiveness of using local modelling techniques to map xate abundances and reveal local abundance hotspots. It highlighted areas needing further survey work and the need for collecting more environmental data to improve the prediction of plant distribution in this region using niche and/or co-kriging modelling approaches. Our study suggested that the xate resource in Belize needs a careful conservation approach.

• 出版日期2009-3