ABOVEGROUND TREE BIOMASS AND LAND USE LAND COVER CHANGE IN BURU COMMUNITY FOREST, TARABA STATE, NIGERIA

Abstract

Remote Sensing (RS) techniques are widely used to estimate Aboveground Tree Biomass (ATB) and for Land Use Land Cover (LULC) classification. Information on ATB and land use changes is essential for development of management strategies for forest ecosystems. Outputs from biomass assessment and LULC classification are constrained by the quality and time of remotely sensed data acquisition. In Nigeria, there is limited information on suitable months for RS data acquisition for estimating ATB and LULC. Therefore, this study was designed to determine the suitable month for RS data acquisition for ATB estimation and LULC classification in Buru Community Forest (BCF), Taraba State, Nigeria.

Landsat imageries of BCF for April, July and December in 1988, 2000, 2008 and 2018 were obtained, based on availability. Twenty (50 m x 50 m) plots were demarcated in BCF and their coordinates were obtained. In each plot, trees with Diameter at Breast Height (DBH, cm) ≥ 5.0 were enumerated and wood core samples obtained. Landsat imageries were classified into LULC. The LULC Changes (LULCC, %) were estimated and projected from 2018 to 2048 using standard methods. Probability (%) of the classified LULCC to remain unchanged from 2018 till 2048 was determined. The DBH and Total Height (TH, m) of trees were measured, while Wood Density (WD, g/cm3), stem volume (m3) and ATB (t/ha) were calculated following standard procedures. Spectral bands of imageries from each month were extracted and used to estimate RS-ATB (t/ha). Suitable month for RS-ATB estimation was selected using highest adjusted coefficient of determination (R_adj^2), Root Mean Square Error (RMSE), Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). The ATB was compared with RS-ATB for 2018 following standard procedures. Data were analysed using descriptive statistics, T-test and linear regression at α0.05.

Six LULC were identified: Less Disturbed Forest (LDF), Disturbed Forest (DF), Farmland, Water Body (WB), Bare Land (BL) and Built-up Area (BA). The LULCC was highest and least in DF (68.70%) and BA (5.13%), respectively. Projected LULCC were: 51.99 (LDF), 31.08 (DF), 12.28 (farmland), 1.65 (WB), 2.58 (BL), and 0.43 (BA). Probability matrix varied from 9.50% (BA) to 69.90% (DF). This suggested that there was a high probability for DF to remain unchanged by 2048. Tree DBH, TH and WD were: 22.56±0.35, 12.86±0.19 and 0.47±0.01, respectively. Stem volume and ATB were 414.66±12.75 and 281.30±0.33, respectively. The highest RS-ATB were 271.66 (1988), 196.60 (2000), 174.50 (2008) and 152.80 (2018) for July (RMSE=444.12, R2adj=94.94%, AIC=246.59, and BIC=21.02), while the least were 148.70 (1988), 146.89 (2000), 122.84 (2008) and 152.60 (2018) for April (RMSE=522.31, R2adj=93.00, AIC=253.08, and BIC=21.34). Estimated ATB (1923.60±101.78) did not differ significantly (t=0.89) from RS-ATB (1910.00±65.67). This implied that RS technique was suitable for aboveground tree biomass estimation in BCF.

Imageries from the month of July were the most suitable remotely sensed data for estimation of aboveground tree biomass in Buru Community Forest, Taraba State, Nigeria. Remote sensing techniques sufficiently predicted aboveground tree biomass and provided accurate land use land cover classification for the forest.