Distribution and habitat preferences of Lecythidaceae in the RÃo Los Amigos uplands, southwestern Amazonia, Peru
2007
Thesis
Master's t
84
Georg-August University of Göttingen
Göttingen, Germany
Los Amigos;plants;angiosperms;Lecythidaceae;trees;distribution;intermediate spatial scales;soils;beta-diversity
Several studies in tropical lowland forests have pointed out the importance of soil properties on plant species distribution patterns. Few studies have analyzed the distribution patterns of canopy tree species and even fewer related those to soil chemical properties. The present study was carried out at two research stations, CICRA and CM2, (200 – 300 m asl.) in the uplands (i.e. ‘terra firme’) of the Los Amigos watershed, Madre de Dios, Peru. The habitat preferences and distribution of the species of the tree family Lecythidaceae were analyzed with respect to chemical properties of top soils. This study addressed the following questions: (i) How is the Lecythidaceae family distributed spatially in the Los Amigos uplands? (ii) How are the individual species of Lecythidaceae distributed spatially in the Los Amigos uplands? (iii) Do soil properties influence the spatial distribution of the species? (iv) To what extent do distance-dependent population processes such as dispersal limitation explain the spatial patterns of the species? A total of 506 individuals of Lecythidaceae out of 11 species (5 genera) were found in 51 transects of 500 m x 20 m (1 ha) each. One soil sample was obtained as a mixed sample from two soil samples taken within each transect from the upper mineral soil (0 – 10 cm) and analyzed chemically. The distribution patterns of the species were analyzed using the ?2-test in order to reveal aggregations of individuals. The relationships between Lecythidaceae composition (at family and species level) and soils were assessed using simple and partial Mantel tests on the basis of dissimilarity matrices. Subsequently, multiple regression analyses were applied on the same basis to characterize the relationships between single species and soil properties. An overview of the abundance and biometric parameters of the encountered species is provided as well as a statistical and descriptive analysis of soil properties of both sites. Results showed that the species Cariniana decandra and Gustavia augusta occurred clumped at both sites whereas Eschweilera tessmannii and Bertholletia excelsa were clumped only in CICRA and CM2, respectively. Soil properties varied substantially and Mantel tests confirmed that Lecythidaceae (density per species / abundance) were statistically significantly correlated with pH (r = 0.19 / 0.24), total phosphorous (Ptot; r = 0.22 / 0.20), aluminum (Al; r = 0.27 / 0.20), and manganese (Mn; r = 0.42 / 0.43) content of the soils in CICRA. The Lecythidaceae abundance also correlated with iron (Fe; r = 0.22) in CICRA. At the species level, the correlation of B. excelsa density was statistically significant with Al at both study sites (r = 0.27 and 0.13, respectively) whereas it correlated with pH (r = 0.23) and Ptot (r = 0.20) in CICRA, and with carbon (C; r = 0.12) inCM2. Gustavia augusta density correlated significantly with C/N ratio (r = 0.09) in CICRA and with pHH2O/KCl and Fe in CM2 (r = 0.18, 0.1 and 0.14, respectively). Eschweilera coriacea density correlated with Ptot, Al and Mn content (r = 0.22, 0.21, and 0.41, respectively) only in CICRA where also E. tessmannii density correlated with nitrogen, C and Fe content (r = 0.11, 0.16, and 0.13, respectively). Cariniana decandra density correlated with C and Fe (r = 0.14 and 0.20, respectively) only in CM2. Multiple and simple regression analyses identified the significant correlations of B. excelsa with pHH2O to be positive and with Al to be negative. The positive correlations of G. augusta with C/N in CICRA and with pHH2O in CM2 were also shown. The positive correlation of E. coriacea with Al and the negative one with Mn was also endorsed for CICRA. Positive correlations were also confirmed for E. tessmannii with C in CICRA and for C. decandra with C and Fe in CM2. It is concluded that Lecythidaceae respond to the variation in soils. Indirectly, this endorses edaphic patchiness as an important determinant of habitat specialization and hence of the spatial distribution of Lecythidaceae species. However, the distribution of each species was found to correspond best to a set of different soil parameters. This implies that the influence of soils is species specific rather than family specific.