Allometric relationship and biomass partition of Kandelia obovata Liu & Yong planted in Nam Dinh province

JOURNAL OF SCIENCE OF HNUE Chemical and Biological Sci., 2013, Vol. 58, No. 9, pp. 91-103 This paper is available online at ALLOMETRIC RELATIONSHIP AND BIOMASS PARTITION OF Kandelia obovata LIU & YONG PLANTED IN NAM DINH PROVINCE Pham Hong Tinh1 and Mai Sy Tuan2 1Mangrove Ecosystem Research Center, Hanoi National University of Education 2Faculty of Biology, Hanoi National University of Education Abstract. Measurements were made of 101 Kandelia obovata Liu & Yong trees growing in Giao Lac Commune, Giao Thuy District, Nam Dinh Province, and, based on the diameter (cm) at 30 cm above the widening of the trunk base, allometric regressions were done to estimate the biomass (kg) of the entire tree, above-ground biomass, trunk biomass, leaf biomass and below-ground biomass, found to be W = 0.10316D1.85845(R2 = 0, 86),W = 0.09012D1.78752(R2 = 0, 84),W = 0.04975D1.94748(R2 = 0, 79),W = 0.00899D1.7643(R2 = 0, 63) and W = 0.01420D2.12146(R2 = 0, 73), respectively. Compared to regressions obtained using the traditional logarithmic transformation method, the above regressions had a much better fit. The original biomass data of the sample trees were also used to calculate the proportion of total biomass apportioned to trunk, branches, leaves and roots. While for small trees (D ≤ 4 cm), 20% is leaves, 20% is branches, 35% is trunk and 25% is roots, for larger trees (D ≥ 10 cm), only about 5% is leaves, 5% is branches, 65% is trunk and 25% is roots. The allometric equations and biomass partition obtained in this study could be used effectively to account for biomass and carbon in Kandelia obovata Liu & Yong mangroves. Keywords: Allometric relationship, Kandelia obovata Liu & Yong, Biomass, Vietnam. 1. Introduction Mangrove forest is an important ecosystem in tropical and subtropical coastal regions. However, this ecosystem is being degraded and damaged at an alarming rate by human activites, e.g. land use change, unsustainable exploitation and human population increase [6]. Moreover, the value of mangrove forests in terms of carbon accumulation and other factors are not well understood. Received September 26, 2013. Accepted December 18, 2013. Contact Pham Hong Tinh, e-mail address: phamtinhsp@yahoo.com 91 Pham Hong Tinh and Mai Sy Tuan Kandelia obovata Liu & Yong is a key species that has been planted in coastal areas of Northern Vietnam, mainly Quang Ninh, Nam Dinh and Thai Binh Provinces. In a natural development and ecological succession of mangrove forest, this species and others grow together in national mangrove forests [2]. Therefore understandings ecological characteristics of this species such as biomass and carbon accumulation could significantly contribute to sustainable management of the mangrove forests. Scientists have developed a number of methods to estimate biomass (dry weight) of both inland and mangrove forest. These methods are divided into three categories: 1) the harvest method, 2) the mean-tree method and 3) the allometric method. The harvest method is not practical for measuring mature forests of the large amount of time and manpower needed. Moreover, the harvest method is not reproducible because the trees are destroyed. The mean-tree method is utilized only in forests that are relative homogeneous in terms of tree size. The allometric method uses allometic equations to estimate the whole and partial weight of a tree referring to measured tree dimensions, including trunk diameter and height. This method is useful for estimating temporal changes in forest biomass by subsequent measurements (Komiyama et al.) [5]. However, allometric equations are site- and species-specific. Little research done using allometric equations to determine Kandelia obovata Liu & Yong biomass has been published or cited. The research presented in this paper is one of the first studies in Vietnam that makes use of allometric equations to determine the total and component biomass of Kandelia obovata Liu & Yong. In addition, we compared two ways of building allometric equations, i.e. determining parameters of equations. In term of carbon accumulation, the roots of mangrove trees are probably the most important component. However, an understanding of carbon accumulation in roots as well as in other components is limited because very few studies have been done on biomass proportion of mangrove trees. Ong et al. [7] have shown that in 20-year-old Rhizophora apiculata mangrove trees, carbon is accumulated in the roots at a rate of 0.42 tC/ha/year, whereas accumulation in the canopy is at a rate of 0.52 tC/ha/year. We have not found published data on carbon accumulation and biomass proportion of Kandelia obovata Liu & Yong. Therefore, in this paper we are the first to announce the biomass proportion of Kandelia obovata Liu & Yong in Vietnam. The results of this study are expected to significantly contribute to biomass and carbon accounting for Kandelia obovata Liu & Yong mangroves in Nam Dinh Province as well as other areas of Northern Vietnam. 2. Content 2.1. Methodology 2.1.1. Study site The present study was carried out onKandelia obovata Liu &Yong that was planted 10 - 20 years ago in Giao Lac Commune, Giao Thuy District, in Nam Dinh Province. Giao 92 Allometric relationship and biomass partition of Kandelia obovata... Lac is located at 20013’-20015’ North, 106015’-106030’ East (Figure 1) and is one of the buffer zones of Xuan Thuy National Park. It is the first Ramsar Site in Vietnam. In Giao Lac there are about 407.7 hectares of mangrove forest. The Giao Lac mangrove trees received a lot of sediment from the Ninh Co and Red River and it is a relatively flat area with a thick layer of alluvial sediment. Mangrove plantations may also be grown in clay mud and sand. Giao Lac has a diurnal tide with sea level of 0.1 - 3.9 m, a temperature of 240C, rainfall of nearly 1500 mm/year, air humidity of about 82% and salinity 18.0 - 28.3%. Figure 1. Location and map of the study site 2.1.2. Materials A sampling of 101 trees was chosen in the study site having diameters ranging from 0.5 to 15 cm. The trees were cut down and the roots dug out in 2008, 2009 and 2013 to measure trunk diameter, total and component dry weight (above ground, stem, branch, leaf and below ground). All sample trees was selected using the stratified sampling method: 1) divide the study area into sections corresponding to the year the trees were planted (from 1998 to 2009); 2) randomly select two or three sample plots (10 × 10 m) within each compartment (a total of 34 sample plots were established) and 3) select three sample trees (smallest, largest and average) within each sample plot. Before felling the sample trees, we measured and recorded the diameter at 30 cm above the widening at the base of the trunk. The sample trees were cut and separated into four components: stems, branches, leaves and roots. Then we weighed and recorded the fresh weight of each component. We took about 5 - 10 g samples of each component, and weighed and recorded the exact fresh weight of each sample. The samples were then taken to the laboratory and dried to a constant weight. The dry samples were weighed again to get the dry weight of each sample. The ratio of fresh weight to dry weight of each sample was used to calculate the dry weight of each component. The total dry weight of the stems, branches and leaves is the above ground biomass. The total biomass of a tree is the sum of the dry weights of all four components. The diameter and dry weight of the sample trees are presented in Table 1. 93 Pham Hong Tinh and Mai Sy Tuan Table 1. Diameter and dry weight of sample trees No. Diamter D (cm) Dried weight W (kg) Total Above - ground Trunk Branches Leaves Below - ground (roots) 1 0.7 0.02 0.02 0.01 0.01 0.01 2 0.7 0.55 0.51 0.01 0.46 0.05 0.04 3 0.7 0.03 0.02 0.01 0.01 0.01 4 0.7 0.08 0.04 0.01 0.04 0.04 5 0.8 0.02 0.01 0.01 0.00 0.01 6 0.8 0.02 0.02 0.01 0.00 0.01 7 1.5 0.02 0.02 0.01 0.01 0.01 8 2.7 1.41 0.83 0.19 0.03 0.60 0.58 9 2.7 0.23 0.13 0.08 0.04 0.10 10 2.9 0.67 0.59 0.10 0.43 0.05 0.08 11 3.1 0.30 0.21 0.12 0.01 0.08 0.09 12 3.1 0.15 0.14 0.14 0.00 0.01 13 3.2 0.19 0.18 0.17 0.01 0.01 14 3.3 0.29 0.24 0.19 0.02 0.03 0.05 15 3.5 0.39 0.29 0.17 0.07 0.05 0.10 16 3.7 0.49 0.32 0.21 0.03 0.08 0.17 17 3.8 1.24 0.73 0.51 0.04 0.18 0.47 18 3.8 1.21 0.76 0.33 0.01 0.42 0.44 19 3.8 0.99 0.90 0.28 0.59 0.04 0.09 20 3.9 0.92 0.75 0.23 0.29 0.23 0.17 21 4.1 1.68 1.14 0.59 0.23 0.32 0.52 22 4.3 0.87 0.77 0.45 0.28 0.05 0.10 23 4.4 1.21 0.90 0.51 0.32 0.07 0.30 24 4.4 1.58 1.13 0.62 0.36 0.14 0.43 25 4.6 2.13 1.60 0.88 0.45 0.27 0.50 26 4.8 3.07 1.55 0.81 0.40 0.34 1.49 27 5.2 1.76 1.24 0.70 0.32 0.22 0.49 28 5.2 2.85 1.61 0.93 0.49 0.19 1.21 29 5.2 1.17 1.11 1.08 0.04 0.06 30 5.3 2.45 1.74 1.40 0.18 0.16 0.69 31 5.3 3.19 1.72 1.03 0.15 0.54 1.43 32 5.3 3.87 3.27 2.64 0.33 0.30 0.57 33 5.3 0.57 0.47 0.33 0.04 0.10 0.10 34 5.4 2.78 2.15 1.81 0.16 0.18 0.60 35 5.4 2.14 1.70 1.44 0.13 0.13 0.41 36 5.6 3.25 2.37 1.90 0.18 0.29 0.87 37 5.6 2.25 1.79 1.01 0.60 0.19 0.44 38 5.6 2.92 2.20 1.01 0.78 0.40 0.71 39 5.7 1.92 1.57 1.08 0.29 0.20 0.35 40 5.7 3.46 2.58 1.66 0.50 0.41 0.82 41 5.7 2.81 2.28 1.15 0.76 0.37 0.53 42 5.7 2.43 2.00 1.51 0.36 0.13 0.38 43 5.7 3.15 2.22 1.41 0.58 0.23 0.91 94 Allometric relationship and biomass partition of Kandelia obovata... 44 5.7 0.63 0.43 0.25 0.04 0.14 0.20 45 5.9 4.33 2.48 1.64 0.24 0.60 1.84 46 6.0 3.43 2.89 2.08 0.65 0.16 0.49 47 6.0 2.23 1.84 1.49 0.20 0.16 0.37 48 6.0 2.46 1.93 1.62 0.06 0.25 0.52 49 6.0 2.92 1.86 1.17 0.19 0.51 1.02 50 6.0 1.00 0.73 0.51 0.08 0.14 0.28 51 6.1 3.37 2.69 2.26 0.09 0.33 0.62 52 6.1 2.37 1.84 1.36 0.27 0.21 0.52 53 6.1 2.98 2.28 1.24 0.50 0.54 0.64 54 6.4 3.32 2.90 2.33 0.30 0.28 0.42 55 6.4 3.02 2.48 2.03 0.15 0.30 0.52 56 6.4 3.78 2.31 1.35 0.40 0.56 1.45 57 6.4 3.49 2.32 2.09 0.23 1.18 58 6.4 3.03 1.80 1.58 0.21 1.23 59 6.5 3.24 2.33 1.69 0.30 0.34 0.90 60 6.7 2.92 1.90 1.59 0.16 0.16 0.98 61 6.7 2.45 2.07 1.60 0.22 0.24 0.38 62 6.7 2.83 2.13 1.79 0.02 0.32 0.71 63 6.7 2.78 2.24 2.19 0.05 0.54 64 6.8 4.09 3.14 2.13 0.73 0.28 0.94 65 7.0 3.37 2.61 2.06 0.25 0.30 0.70 66 7.0 3.45 2.45 1.97 0.22 0.26 0.99 67 7.0 3.38 2.08 1.82 0.27 1.30 68 7.1 2.64 2.16 1.70 0.24 0.22 0.46 69 7.2 3.91 2.65 1.68 0.63 0.34 1.24 70 7.2 1.69 1.40 1.18 0.06 0.16 0.22 71 7.3 4.97 3.61 2.77 0.47 0.37 1.36 72 7.5 4.75 3.92 3.13 0.24 0.54 0.82 73 7.7 4.28 3.49 3.05 0.16 0.27 0.78 74 7.8 3.39 2.58 2.45 0.13 0.80 75 7.9 5.05 2.76 2.49 0.09 0.18 2.26 76 7.9 4.89 4.22 3.72 0.17 0.32 0.65 77 8.0 4.66 3.21 2.40 0.25 0.55 1.37 78 8.0 3.57 1.83 1.16 0.36 0.32 1.68 79 8.0 2.32 1.99 1.66 0.02 0.31 0.33 80 8.3 5.94 5.01 4.61 0.24 0.16 0.81 81 8.3 6.04 4.52 3.97 0.24 0.31 1.48 82 8.4 4.65 3.90 3.08 0.41 0.41 0.73 83 8.6 7.99 6.61 5.36 0.74 0.50 1.30 84 8.6 3.20 2.31 1.62 0.40 0.30 0.82 85 8.6 2.51 1.73 1.22 0.31 0.21 0.77 86 8.8 7.82 6.37 5.16 0.74 0.47 1.37 87 8.8 7.46 6.26 5.51 0.28 0.47 1.12 88 8.8 7.56 6.44 5.51 0.46 0.47 1.12 89 8.9 7.87 6.67 5.61 0.31 0.75 1.16 90 9.0 5.03 4.55 4.14 0.28 0.13 0.46 95 Pham Hong Tinh and Mai Sy Tuan 91 9.1 8.28 6.76 5.80 0.46 0.51 1.48 92 9.2 8.46 5.90 5.29 0.30 0.32 2.52 93 9.2 6.57 5.25 4.64 0.13 0.48 1.31 94 9.3 4.74 3.54 2.91 0.28 0.35 1.20 95 9.4 3.47 2.76 2.32 0.27 0.18 0.67 96 9.6 5.52 4.23 3.52 0.31 0.40 1.22 97 10.5 9.73 5.94 5.39 0.26 0.29 3.79 98 12.1 12.64 7.90 7.07 0.34 0.49 4.72 99 13.0 11.22 9.61 8.39 0.25 0.96 1.59 100 13.7 13.33 8.44 7.52 0.34 0.59 4.85 101 13.7 8.66 6.50 5.37 0.53 0.60 2.14 2.1.3. Allometry and variables of allometric equations Allometry is based on the fact that there is proportionality between the relative growth rates of two different parts of the plant [3]. The relationship between the two variables can be expressed by the following equation: ( dy ydt ) : ( dx xdt ) = k (1) where x is the independent variable (e.g. stem diameter, tree height or both), y is the dependent variable (e.g. biomass) and b and k are the allometric constants. The equation can be simplified as follows: (1)⇔ dy y = k dx x ⇔ ∫ dy y = k ∫ dx x ⇔ ln y + c1 = k (lnx+ c2) ⇔ ln y = ln xk + ln b ⇔ y = bxk (2) In this study, we used diameter at 0.3 m height (D) as the independent variable because it is easy to measure in the field by using a caliper or tape measure. Tree height was not used in the equation because measuring the height of mangrove trees does not in an accurate and it is time consuming. In the latest guidelines on carbon assessment of mangroves, CIFOR (Center for International Forestry Research) recommended that the allometric equation should be developed and used with one independent variable – the diameter [4]. 2.1.4. Allometric constant determination As the allometric equation form and variables of equations were determined, we used the statistical analysis software JMPIN with the input data presented in Table 1 to 96 Allometric relationship and biomass partition of Kandelia obovata... find the allometric constants b and k. The equations obtained by JMPIN were compared with the corresponding equations developed using the traditional approach. For the traditional method, the equation (2) was represented on logarithmic coordinates as a linear relationship of the form: lny = klnx + b (3). The measured date (Table 1) was also converted from arithmatic to logarithmic form in order to correspond with the equation (3). Then the least squares method was used to find the equation parameters b and k. 2.1.5. Allometric equation evaluation The obtained allometric equations were evaluated using the the coefficient of determination (R2), sum squared error of estimate (SE) and mean error of estimate (bias) of the allometric equations developed by both approaches. Bias, R2 and SEE were calculated by equation (4), (5) and (6) respectively. R2 is a statistic that would give information about the goodness of fit of the allometric equations (i.e. R2 is a statistical measure of how well the regression curve/line approximates the real data points). An R2 of 1 indicates that the regression curves/lines fit the data perfectly. SSE and bias are errors of estimates when using allometric equations compared with the true values. R2 = 1−   N∑ i=1 (yi − y , i) 2 N∑ i=1 ( yi − 1 N N∑ i=1 yi )2   (4) SSE = N∑ i=1 (yi − y , i) 2 (5) bias = N∑ i=1 (yi − y , i) N (6) where yi is biomass of i th tree measured in the field, y′i is biomass of i th tree using the allometric equation and N is total number of sample trees. 2.2. Results and discussions 2.2.1. Total biomass allometry Table 2. Allometric equations for total biomass Log - transformed Equation D (cm) N R2 SSE bias Yes lnW = 2.31487lnD – 3.26991 (W = 0.03801D2.31487) 0 – 15 101 0.82 154.49 0.13 No W = 0.10316D1.85845 0 – 15 101 0.86 113.43 -0.07 97 Pham Hong Tinh and Mai Sy Tuan Table 2 presents the allometric regressions obtained for total biomass with log-transformation and non-transformation. The coefficients of determination (R2) for both equations are greater than 0.8. They indicate that a large proportion (over 80%) of the total biomass could be determined when referring to a diameter at a height of 0.3 m. However, the errors of estimates (SE and bias) show that the non-transformed equation fits the data better than the log-transformed equation because their SSEs are 113.43 and 154.49 respectively, and their biases are -0.07 and 0.13 respectively. Moreover, the non-transformed and log-transformed regressions fitting the actual data shown in Figure 2 also indicate that the fit was better with the non-transformed data than with the log-transformed data. For trees with a diameter less than 5 cm, two equations overestimate the actual biomass. For trees with a diameter greater than 10 cm, the log-transformed equation also overestimates the actual biomass, but the non-transformed equation yields a better estimate. Figure 2. Total biomass against diameter of actual and two fitted models Figure 3. Total above-ground biomass against diameter of actual and two fitted models 2.2.2. Above-ground allometry Total above-ground biomass The allometric regressions developed for total above-ground biomass with log-transformation and non-transformation are presented in Table 3. Table 3. Allometric equations for total above-ground biomass Log - transformed Equation D (cm) N R2 SSE bias Yes lnW = 2.33585lnD – 3.59103 (W = 0.02757D2.33585) 0 – 15 101 0.76 111.82 0.09 No W = 0.09012D1.78752 0 – 15 101 0.84 76.28 -0.06 98 Allometric relationship and biomass partition of Kandelia obovata... The coefficient of determination (R2) for a non-transformed equation is 0.84 and for a log-transformed equation it is 0.76. This indicates that their proportions of total above-ground biomass when referring to a diameter at 0.3 m height are 84% and 76% respectively. On the other hand, SSE and bias for the log-transformed equation are 111.82 and 0.09 while the values for the non-transformed equation are 76.28 and -0.06, respectively. Thus, it is clear that the non-transformed equation fits the data better than the log-transformed equation. A detailed consideration of the graphs of total above-ground biomass against the diameter of the actual and two fitted models (Figure 3) also indicates that for trees with a diameter greater than 5 cm, the non-transformed equation fits the data better than the other. Trunk biomass The allometric regressions developed for trunk biomass with log-transformation and non-transformation are presented in Table 4. Table 4. Allometric equations for trunk biomass Log - transformed Equation D (cm) N R2 SSE bias Yes lnW = 2.22714lnD – 4.26584 (W = 0.01404D2.22714) 0 - 15 101 0.76 67.66 0.19 No W = 0.02485D2.04924 0 - 15 101 0.80 60.83 -0.04 Figure 4. Trunk biomass versus diameter of actual and two fitted models Figure 5. Branch biomass versus diameter of actual and two fitted models Both equations have a coefficient of determination (R2) of around 80% indicating that about 80% of the trunk biomass could be determined referring to the diameter. However, SSE and bias show that the non-transformed equation fits the actual data better than the log-transformed equations because SSE and bias for the log-transformed equation are 67.66 and 0.19 while the values for non-transformed equation are 60.83 and -0.04, respectively. The graphs of trunk biomass against diameter of the actual and two fitted models (Figure 4) also indicate this for trees with a diameter that is greater than 10 cm. For 99 Pham Hong Tinh and Mai Sy Tuan trees with a diameter less than 10 cm, both equations have an almost equivalent goodness of fit. Branch biomass Branches were not separated from the total biomass in 15 of the 101 trees sampled. The allometric regressions developed for branch biomass with log-transformation and non-transformation are presented in Table 5. Table 5. Allometric equations for branch biomass Log - transformed Equation D (cm) N R2 SSE bias Yes lnW = 1.35289lnD – 3.17169 (W = 0.02429D1.35289) 0 - 15 86 0.57 2.56 1.40 No W = 0.00209D2.56871 0 - 15 86 0.73 1.83 -0.38 The calculations of coefficient of determination (R2) and errors of estimates (SSE and bias) indicate that the non-transformed equation fits the actual data much better than the log-transformed equation. The values for the non-transformed equation are 0.73, 1, 83 and -0.38, respectively, while the values for the log-transformed equation are 0.57, 2.56 and 1.40, respectively. The graph of branch biomass against diameter of the actual and two fitted models (Figure 5) also clearly shows that the log-transformed equation does