Development of regression models for predicting yield of <i>Triplochiton scleroxylon</i> (K. Schum) stand in Onigambari Forest Reserve, Oyo State, Nigeria

Abstract

We developed regression models for predicting yield of Triplochiton scleroxylon stands in Onigambari Forest Reserve in western Nigeria. Stratified random sampling technique was adopted for separating the stands into series based on their ages, viz.: 1979 plantation (40-year-old), 1988 plantation (31-year-old) and 1992 (27-year-old). Seventy five (75) 20m × 20m (0.04 ha) temporary sample plots were used (i.e. twenty five in each age series). Tree growth variables including diameter at breast height (Dbh); total height (THT); merchantable height (MHT); stem quality (SQ); crown length (CL) and crown diameter (CD) were measured on all trees within the sample plots to compute basal area (BA), slenderness coefficient (SC) and stem volume (SV). Trees were classified according to their slenderness coefficient (SC) as high (SC: >80), moderate (SC: 70-80) and low (SC: <70). Data were analyzed using descriptive statistics and regression analysis. Fitted models were assessed using F-test and coefficient of determination (R²). Model validations were conducted using t-test and bias. The result showed that mean Dbh for 40, 31 and 27-year-old stands were 67.99 ± 15.40 cm, 50.97 ± 9.71 cm and 36.85 ± 5.19 cm, respectively, and the overall mean Dbh was 55.67 cm which was higher than 48 cm stipulated as the minimum merchantable size by logging policy of Nigeria. The diameter class >50 cm had the highest of 564 trees/ha followed by 40-50 cm diameter class with 317 trees/ha. The diameter class with the least number of trees was 10-20 cm with 4 trees/ha. The mean tree height for 40, 31 and 27-year-old were 36.97 ± 6.84 m, 34.84 ± 6.99 m and 25.38 ± 5.93 m, respectively. Height class 30-40 m recorded the highest number of 489 trees/ha followed by 20-30 m with 405 trees/ha while >50 m height class recorded the least number of 12 trees/ha. There were 953 trees/ha (78.44%) with low to moderate slenderness coefficient indicating that most of the trees in the study area were of good vigour, and could withstand windthrow or other wind-induced damages. The best yield prediction model was of the form: V= -0.123 + 0.00673Dbh³ – 3.419BA³ + 0.020934MHT³ with R², SSE and bias values of 98.2%, 0.644 and 0.102, respectively. It is therefore recommended for future yield prediction in the study area.

Keywords: Stand, yield prediction, stem diameter, height classes, age series