Decision support tools for the management in a Dry Afromontane Forest in Ethiopia

Abstract

Ethiopia is one of the tropical countries endowed with diverse forest formations. These forests provide large amounts of wood that can be used for furniture, construction, and domestic energy consumption. However, the gap between the supply and demand of wood is huge and increasing. This growing demand could be met by the sustainable production of wood from the existing natural forests. This requires up-to-date information on forest structure, diameter growth rate, standing volume, cutting cycle, and minimum logging diameter of trees. Unfortunately, such information is not available for trees from the Chilimo Dry Afromontane Forest in Ethiopia. To address this gap, this study aimed to i) develop a nonlinear mixed-effects model for predicting Juniperus procera tree height (chapter one); ii) develop mixed-species allometric equations to quantify stem volume and biomass of trees from the Chilimo Dry Afromontane Forest (chapter two), and iii) ensuring sustainable wood harvesting from J. procera trees in the Chilimo Dry Afromontane Forest (chapter three). For the first chapter, a total of 1,215 height and diameter measurements were recorded on 101 sample plots. The best- fitting base model was selected after a comparison of fourteen models. We also included the sample plot as a random effect in non-linear mixed effect modeling. The effect of adding stand variables on height prediction performance was also evaluated. The bias, root mean square error, and AIC were computed and used as the model evaluation criteria. We found that the Michaelis-Menten model best represented the height-diameter allometry of J. procera trees. The best mixed-effects model (M1) improved the height prediction performance with the RMSE and bias values of 2.692 and 0.043, respectively. The addition of the quadratic mean diameter and stem density slightly improved the prediction performance of the best-mixed effects model. The calibration response revealed that the systematic selection of the three largest diameter trees in a sample plot is the best sampling alternative to estimate the random effects and predict the height of J. procera trees from the new plots or stands. For the second chapter, we used a total of 194 sample trees from seven dominant tree species (Juniperus procera, Podocarpus falcatus, Allophylus abyssinicus, Olea africana ssp. Cuspidata, Olinia rochetiana, Rhus glutinosa, and Scolopia theifolia). Various volume and biomass equations were fitted using robust linear and nonlinear regression models. Model comparison indicated that the best model to estimate stem volume was 𝑙𝑛(𝑣)=−9.909 + 0.954 𝑙𝑛 ( 𝑑𝑏ℎ 2 ℎ𝑡 ) , whereas the best model to estimate biomass was 𝑙𝑛(𝑏) =−2.983 + 0.949 𝑙𝑛 ( 𝜌𝑑𝑏ℎ 2 ℎ𝑡 ) . These equations explained over 85 % of the variations in the stem volume and biomass measurements. The mean density and basal area of trees with dbh ≥ 2 cm were 24.4 m 2 ha -1 and 631.5 stems ha -1 , respectively. Based on the newly developed equations, the forest has on average 303.0 m 3 ha -1 standing volume of wood and 283.8 Mg ha -1 biomass stock. The newly developed allometric equations derived from this study can be used to accurately determine the stem volume, biomass, and carbon storage in the Afromontane forests in Ethiopia and elsewhere with similar stand characteristics and ecological conditions. By contrast, the generic pan-tropical and other local models appear to provide biased estimates and are less appropriate for dry Afromontane forests in Ethiopia. For the third chapter, we established 165 plots (each 400 m 2 ) in the forest and collected vegetation data. We also conducted growth ring measurements on 12-disc samples from J. procera trees. We determined the diameter growth rate, the current and mean annual increments, the minimum logging diameter, and the cutting cycle. By using the stand projection table, we estimated the harvestable volume of wood by combining four minimum logging diameters and five cutting cycles. The findings revealed that J. procera tree species has a mean density of 183 stems ha -1 , a total basal area of 12.1 m 2 ha -1 , and 98.9 m 3 ha -1 standing volume of wood. The population exhibited an inverted J-shape diameter distribution pattern. The mean annual diameter growth rate ranges between 0.50 and 0.65 cm yr -1 , with an overall mean of 0.59 cm yr -1 . The current annual increment occurred at 50 years when trees reached 30 cm in diameter, while the mean annual increment occurred at 90 years when trees attained 50 cm in diameter. After evaluating various scenarios, we found that a minimum logging diameter of 40 cm and a cutting cycle of 15 years provided the highest harvestable volume of wood (22 m 3 ha -1 ) and volume increments (1.4 m 3 ha -1 yr -1 ). Additionally, this scenario allows for the harvesting of 9% of the standing J. procera trees while maintaining a larger proportion (91%) of the existing standing trees in the forest. Based on our findings, we concluded that the Chilimo Dry Afromontane Forest is well stocked and has a substantial amount of harvestable wood volume, which could help Ethiopia meet its growing national wood demand. Our study provides valuable information for policy makers to formulate regulations for wood harvesting from J. procera trees in Chilimo forest. Keywords: Wood production. Forest structure. Allometeric equations. Dendrochronology. Mixed models.