These analyses were done considering three key variables, i.e. gear, habitat see more (where fishing took place) and time (northeast monsoon, dry, southeast monsoon). Two 3-way ANOVAs with the above variables and their respective interactions were performed; one for biomass and one for income (Appendix III, Supplementary Information). When significant differences occurred (p < 0.05) the Bonferroni correction (BC) was applied to determine the final significant differences between habitats. For each ANOVA pairwise tests were performed summing up to 72 pairwise tests totally ( Appendix III, Supplementary Information). The significance level for the pairwise tests was determined by the critical p-value based
on the BC, i.e. 0.05/36 = 0.00139. To better fulfill the ANOVA assumptions on normality and variance homogeneity the analysis was performed on log-transformed values. All the statistical analyses were performed with the statistical program
Stata version 12. Fish species composition was calculated using the relative abundance of the species found in each “batch” brought to the market belonging to the selected three habitats, i.e. mangroves, seagrasses and corals. GW-572016 in vitro Data was then aggregated by time (season) and pooled for all habitats to determine the most common species found in the bay. This analysis, although lacking details, provides a clear indication of what type of fish dominates the catches in Chwaka Bay (Table 2). The study limitations are acknowledged in the sense that only the biggest market in the bay was sampled and that there is no replication over time. However, the choice was based on the fact that the Chwaka market is the largest and most important within the bay but also in Zanzibar where seagrass associated fish is very common in catches for the whole Island (DFMR, 2007). Spatial replication is considered acceptable since we are using a case study approach and each area dominated by the particular habitat within the bay was composed of numerous fishing grounds. All these grounds were mapped those and all fish harvested in those
areas was sampled (see above). The restrictions in sampling were due to logistical reasons since sampling in these rural developing areas is highly resource demanding. However, the results are considered reliable and valid enough to illustrate the arguments and to promote better management. The data analysis showed that fishing takes place in the three investigated habitats (mangroves, seagrasses and corals) in Chwaka Bay (Table 1, Fig. 2). However, compared to mangroves and coral dominated fishing grounds, seagrass dominated grounds were the most visited places for fish harvesting (Fig. 2). The dominating gears in the area were basket traps, drag-nets and spears. The fishing pressure (No. fishers km−2 day−1) varied a lot between the three habitats, but with seagrasses showing the highest (Table 1, Fig. 2).