Comparison of measured and modelled growth on permanent plots in Sabahs rain forests Peter Ko Èhler a,* , Thomas Ditzer a , Robert C. Ong b , Andreas Huth a a Center for Environmental Systems Research, University of Kassel, Kurt-Wolters-Str. 3, D-34109 Kassel, Germany b Forest Research Centre, Forestry Department Sabah, P.O. Box 1407, 90008 Sandakan, Sabah, Malaysia Received 15 June 1999; received in revised form 26 January 2000; accepted 6 February 2000 Abstract In this paper previously unpublished ®eld data from 25 ha of permanent sampling plots (PSPs) in Sabah, Malaysia, in four different forest reserves are analysed for mortality rates and basal area development. Field data of an observation length of 9 to 20 years were available. These data then form the basis of several benchmark tests for the evaluation of the individual-oriented tropical rain forest growth model FORMIND. A new version of the FORMIND is presented. The model in its version FORMIND1.1 includes enhanced submodels for mortality and tree growth. The model evaluation is focused on the model components for tree growth, competition and mortality. Data for tree recruitment were not available. Results show a good agreement between simulation and ®eld data for the main output variables basal area and stem number indicating a reasonable behaviour of the model components we focused on. Furthermore the results show that differences in site conditions in¯uence tree growth and mortality. Site characteristics should be included in the model in the future. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Forest growth model; Tropical rain forest; Dipterocarp forest; Mortality; Malaysia; Basal area; FORMIND 1. Introduction Evaluation of forest growth models is an important procedure of model development. Vanclay and Skovsgaard (1997) discussed range and importance of model evaluation. An evaluation of tropical rain forest models is dif®cult due to a lack of adequate ®eld data. Besides a comparison of model output in a steady state with primary rain forest data (e.g. Bossel and Krieger, 1994; Huth et al., 1994, 1998; Ku Èrpick et al., 1997; Ko Èhler and Huth, 1998a, b) few perma- nent sampling plot data exist which are suitable for testing rain forest growth models. The ®eld data used in this paper were not available to the authors in times of model development and can therefore serve for testing of model results. Because those ®eld data are used here for elaborating parameter values of mortality, they cannot be seen as fully independent, but semi-independent. The data used in the following were collected in the forest reserves Garinono, Gunung Rara, Segaliud Lokan and Sepilok in Sabah, Malaysia by the Forestry Department. Analysis in terms of site and stand char- acteristics, mortality and recruitment data are unpub- lished and only available in several research reports (Ong and Kleine, 1995; Ko Èhler, 1998). Details of the ®eld data used here are therefore documented in this paper. Especially mortality rates are analysed as a function of time and different plant functional types (PFTs). Forest Ecology and Management 144 (2001) 101±111 * Corresponding author. Tel.: 49-561-804-3175, fax: 49-561-804-3176. E-mail address: koehler@usf.uni-kassel.de (P. Ko Èhler). 0378-1127/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII:S0378-1127(00)00364-9