MODELLING GROWTH AND YIELD OF PLANTATION-GROWN Cupressus lusitanica AND Pinus patula IN CENTRAL REGION OF KENYA

Abstract

In this study, forest inventory data as well as data from destructive sampling of Cupressus lusitanica and Pinus patula from Central Kenya was used to develop growth and yield models for the two species. The main aim of the study was to update the models currently in use thereby enhancing plantation forest management. Data from forest plantation inventory carried out by Kenya Forest Service between 2009 and 2011 was used to derive growth variable relationships. In addition, destructive sampling was carried out where trees were randomly selected from plantations of different ages in Kiambu and Nyandarua counties. Tree diameters were taken at butt end, 30cm, 60cm, 1m, 1.3m and subsequently at every 1m interval to the tip of the felled tree. This data was used to calculate volume of trees using the Smalian’s formula which enabled comparing the performance of volume models currently in use with the volume models developed in this study for both species respectively. Goodness of fit of the models was determined by use of Coefficient of Determination, Root Mean Square Error and p-values. The study developed a set of growth models based on height against age, height against dbh as well as the dhb against age. Tree height and age showed exponential relationship of the form Ht = b(Age)c, while the height and dbh was developed using models of the form Ht = a + b(dbh)c and dbh was related to the age by equation of the form dbh = a(Age)b. The height and dbh of P. patula followed a linear trend of the form H = a + b(dbh). These results suggest that the new models and parameters were species specific. The relationship between the volume measured by Smalian’s formula and dbh and height of C. lusitanica and P. patula was developed using third order differential equation. The predicted volume of the two tree species showed better model fit with measured volume than the models currently in use (RMSE = 0.0873 against 0.5065) for C lusitanica and (RMSE = 0.0902 against 0.4648) for P.patula respectively. The latter models overestimated volumes at extreme high or low values but underestimated most intermediate volumes for both species.