Demography and potential extractive use of the liana palm, Desmoncus orthacanthos Martius (Arecaceae), in southern Quintana Roo, Mexico

In order to evaluate the extractive use potential of Desmoncus orthacanthos, various matrix models were employed to assess population structure and dynamics. To this end, populations of this liana palm were investigated in southern Quintana Roo, Mexico, at two localities showing contrasting fragmentation levels arising from different land use regimes. In one case continuous forest (CF) are maintained under forestry management, while in the other (fragmented forest, FF) isolated forest patches remain inside of a matrix of crop and livestock lands. Three forest conditions: mature forest (MF), young forest (YF) and forest edge (FE) were nested in each fragmentation level. Results showed that more plants were encountered under FF situations than under CF conditions; however, no differences were noted in exploitable shoots (i.e., total numbers and lengths) when both were compared. Under forest conditions, number and length of exploitable shoots did not differ between FE and MF populations, however these values were smallest in YF populations. Intrinsic population growth rates (lambda) were lower under FF than under CF, as reflected by periodic matrix values of 0.964 and 1.594, respectively. With respect to forest conditions, lambda for FE > lambda for YF > lambda for MF and periodic matrix values were 1.441, 1.193 and 1.075, respectively. Elasticity analyses for annual matrices showed that the demographic process having most influence on variations in lambda was (periodic matrix values) permanence (43-63%), followed by growth and retrogression (15-22% and 14-23%, respectively); fecundity had the lowest influence in changes of lambda-values (8-13%). Analyses of the simulated extraction of shoots greater than or equal to5 m suggested that FE and CF populations might support harvest rates of 40% per year, since lambda > 1; however, simulation of harvest rates as low as 20% in other populations always resulted in lambda < 1. Interestingly, the simulated addition of nursery-raised juveniles resulted in a notable increase in lambda, especially for FE populations; simulations combining shoot extraction with the addition of juveniles suggested that high levels of harvest may be possible, even while maintaining lambda > 1; but only if sound forest management is practiced and agricultural fires are restricted. This work showed the value of matrix models for analyzing populations dynamics in establishing sustainable use strategies for D. orthacanthos as opposed to static evaluation of population numbers. (C) 2003 Elsevier B.V. All rights reserved.