Study area
This study focused on the estuarine mangroves of the Lindi region in Tanzania that currently has approximately 4,500 ha of mangroves, of the 108,000 ha found in Tanzania (Figure 1). Although this areal extent should be seen as an estimate, as the area of mangroves is not fully known and depends on how these are accounted for; for example UNEP-WCMC estimated 127,200 ha, in 2000, Francis and Bryson estimated 133,500 ha in 2001 (38,39).The average annual temperature in Lindi is 25.7 °C, mean annual rainfall is 1200mm yr−1, with a rainy season that extends from October to June (40). The coastal soils in the region consists of alluvial and sandy soils (41). Mangroves of the study area are supported by the presence of Lukuledi, Ngurumahamba, Mtange and Mingoyo rivers mostly from Rondo catchment with the exception of the Lukuledi river which originates from Nachingwea.
Mangrove ecosystems in Tanzania have long been exploited by humans. Before and during the colonial era, poles and timber were used as building materials for boats and houses by Arabic traders (42,43). Mangroves continue to be exploited for firewood and poles, but large timbers requited for boats are no longer available (5,44). In the study area mangroves are used as source of building poles for houses, and fuelwood for lime burning to create cement (44), as well as being cleared to provide space for seaweed cultivation, illegal sand mining and salt pan construction for salt making (45,46). In Lindi mangrove ecosystems, construction of salt pan is mostly conducted close to the shore to allow easy and lower cost feeding of ocean water to the constructed ponds. This is associated with creation of salt pans pathways, salt storage areas and huts construction. These activities have largely been affecting the growth and stocking of mangrove species (46). Similarly, seaweed cultivation is mostly practiced close to the shore where ocean water is permanently available. Farmers will look for open areas or clear mangrove areas, as they do for salt open construction, to establish their farms. This intensifies mangrove degradation close the shore and hence has an implication to carbon storage. This has been the opposite for illegal harvest for timber, building poles and firewood which is mostly practiced away from the shore for easy transport to the desired destinations.
Figure 1 (a) Tanzanian coastline with mangroves highlighted in green; (b) Close-up of Lindi estuarine mangroves with study plots (black circles); (c) Study area in Tanzania; (d) Google satellite image of plots showing the proximity of farming; Mangrove coverage extracted from Bunting et al., (2018) (47).
Data collection and analysis of soil samples
Four 1-ha plots were established at 4.3km, 8.1km, 11km and 13.5km along a gradient through the mangrove forest from the shore to land (Figure 1). 1-ha plots were divided into 20 subplots of 20 x 20m (48,49). These subplots were separated from one another using sisal ropes creating visual borders to avoid double measurements of stems. A systematic pattern (North-South) was then followed to measure stems in each subplot (Figure 2). In each sub plot, the diameter at breast height (DBH; 1.3m), the species and the height were recorded for all stems ≥10.0cm. The same variables were recorded for smaller stems (≥5.0-9.9cm DBH) in five subplots of 20x 20m (subplots 1, 5, 13, 21 and 25). Stem heights for the trees ≤ 10.0m height were measured parallel to tree from the base to the highest point using a pole of known height (50). Heights of the trees >10m were measured using a laser distance meter (Leica disto). For species which were not identified in the field, a voucher specimen was collected and taken to the National Herbarium in Arusha for further identification. In total, we sampled 2071 stems ≥10.0cm and 970 stems ≥5-9.9cm. Seven species of mangrove were found. Given the homogenous nature of the mangrove ecosystem this was deemed to capture the extent of any variation and provide insight into patterns of above and below ground carbon storage.
Figure 2: 1 ha Vegetation plot, showing movement between numbered sub-plots
In each plot, litter biomass was recorded as follows: first, 1m2 quadrats were established in the corners of subplots 1, 5, 21, 25 and at the centre of the sub plot 13. Litter materials (excluding dead wood) were collected from the five (1 m2) established quadrats and the total wet weight was taken. Sub-samples (50%) were taken from the whole sample, weighed before packing and transported to the lab (51,52). The wet combustion method was used to estimate percentage organic carbon from the dry mass of the litter (53). A portion (50%) of the litter was oven dried to constant weight at 70.0◦C to determine the dry mass (54) and grounded to fine powder for total organic carbon determination. The total organic carbon for litter was determined using the wet combustion procedure as described in Nelson & Sommers (55). The amount of carbon in each sample was calculated as the product of percentage organic carbon and dry mass (54).
A pit of 1m depth was dug 15m away from each 1 ha plot. Due to the challenging environment of the mangrove ecosystem, soil pits were allocated in such a way that samples could be collected up to 1m without water interference, through careful timing of the water tides. Soil samples were collected using a metal ring (98.12 cm volume) inserted into the sediment in a pit dug from a profile at different depths: 0-15cm, 16-30cm 31-60cm and 61-100cm. Each layer was packed separately, and soil samples were transported to the lab, air dried, grounded and passed through a 2mm sieve to remove stones and gravel. SOC was determined based on the Walkley-Black chromic acid wet oxidation method (56) and the results were expressed as the % organic carbon. Computation of SOC density was based on soil mass per unit area obtained as the product of soil volume and soil bulk density determined from the bulk density samples in (g/cm3).
AGC estimations and Data Analysis
Above ground biomass of all stems ≥ 5.0cm DBH (AGB, Mg ha-1) was computed using different biomass equations, including generic equations derived by Komiyama et al. (2008) and Chave et al. (2014) (57,58) (see Additional Information). We report here the values of AGB and below ground biomass (BGB) using the multispecies equations developed by Njana et al. (2015), as these equations were derived using species from coastal regions in Tanzania, including Lindi (36). AGC and BGC (Mg C ha-1) stocks were determined by using a carbon fraction of 0.47 and 0.39, respectively (59–61). We computed AGC using stems ≥ 10.0cm (named AGC10), and also using stems ≥ 5.0cm (named AGC). We assessed the intra-plot variation in AGC by randomly sampling smaller areas (400m2, 1600m2, 3600m2 and 6400m2) of each 1ha plot. The standard deviation relative to sampling the full 1ha was calculated using a bootstrapping approach of 10,000 iterations. For each 1ha plot we computed stem density (stems ha-1), percentage of small stems (those 5.0-9.9cm DBH), basal area(in m2 ha-1), mean diameter (cm), mean height (m), species’ richness (number species present in the plot), species’ dominance (in terms of basal area), and species’ contribution to plot-level AGC (in percentage). Statistical analysis was carried out using R Studio (version 3.6.0). Pearson correlation coefficient was used to determine correlation between seaward distance and AGC or SOC. Paired t-tests were used to compare significant differences between AGC and AGC10.
To compare our findings with those reported elsewhere across Africa, we carried out a literature review searching for mangrove carbon estimates across Africa.