Md. Sarwar Hossain
Palaeoecology Laboratory, Geography and Environment, University of Southampton, Southampton SO17 1BJ, UK
John A. Dearing
Palaeoecology Laboratory, Geography and Environment, University of Southampton, Southampton SO17 1BJ, UK
M. M. Rahman
Institute of Water and Flood Management (IWFM), Bangladesh University of Engineering and Technology (BUET), Dhaka 1000, Bangladesh
M. Salehin
Institute of Water and Flood Management (IWFM), Bangladesh University of Engineering and Technology (BUET), Dhaka 1000, Bangladesh
Bangladesh, Delta, Ecosystem services, Well-being, Trend analysis, Tipping points Int
South-west coastal zone of Bangladesh
Socio-economic and Policy
Social status
A study framework based on the ideas that ecosystem services (e.g. water availability) are derived from ecosystem processes (e.g. nutrient cycling, primary production, soil erosion) and that these services have impacts on human well-being. Ecosystem processes are the basis of ecosystem services but are currently omitted here due to lack of data (e.g. primary production). However, ecosystem processes are increasingly discriminated from ecosystem services in order to avoid double counting). Ecosystem services indicators was selected based on the study framework, data availability, measurability and current environmental concerns (e.g. increasing soil salinity).
Food, shrimp, fish and raw materials, including timber, were selected to show the trajectories of provisioning services. People of the delta depend on a variety of crops, such as different varieties of rice, jute and sugarcane. Temperature and precipitation were used as the performance indicators of local climate. Fluvial erosion and floodplain accretion rates are the performance indicators for erosion prevention. It choosen the time series statistics of crops naturally damaged by different disasters (flood, cyclone, water logging etc.) to analyse the trends of natural hazard mitigation. For indicators of biodiversity, use trees per hectare (ha), growing stock and the number of Royal Bengal tigers and deer in the mangrove forest. The mangrove forest covers *28% of the study area with *10 million people directly and indirectly dependent on the ecosystem services it provides. Moreover, the total coastal ecosystem produces around US$277 GDP per capita (Sarwar 2005). Collections of raw materials (mainly forest products) from the mangrove are secondary occupations. However, biodiversity data for other parts of the study area are not available. Missing values amount to 5% for most indicators (e.g. rainfall and water discharge), and these were analysed using statistical test that assumes continuous series. Where time series data were limited to specific years, these are shown only as simple plots.
Continuous series are expressed in raw and normalised z score forms. The data were analysed using linear regression and nonparametric Mann– Kendall statistics (MK stat) and the Lepage test in order to discriminate between improving, deteriorating and stable trends and to identify statistically significant change points. It was applied the nonparametric Mann–Kendall trend test to detect trends in the time series data of water discharges, ground water level, rainfall and temperature. High variability and nonlinear system characteristics are the reasons for applying this nonparametric test. The Mann– Kendall test has already been applied in detecting temperature, precipitation and river flow data. The nonparametric Lepage test was also applied to detect change points in the trajectories of ecosystem services. The Lepage test has been used to detect change point or stepwise change for rainfall, climate change and water resources. Aggregated and smoothed indices are produced from mean z scores in order to compare trends in different sets of indicators. The main rice crop is used as the index for food provisioning services. The trajectory of salinity has been reversed to represent a water quality index. Lowess smoothing has been applied to these entire indices before plotting. Lowess is the process of time series smoothing similar to running mean of about 15 years that removes the high and low frequencies from the time series data (Wilkinson 1997). This smoothing method is already applied to climate-related and water-quality-related studies. It was analysed the controls on the value-added production from agriculture and revenue from the mangrove forest products using bivariate plots, Spearman’s Rank correlation coefficient and the literature review. Trend observations and bivariate plots are used to discriminate between slow and fast variables. It was attempted to identify slow and fast variables of ecosystem services, where the slow variables are the controlling and shaping variables for system resilience (Biggs et al. 2012; Walker et al. 2012). The literature review, bivariate plots and correlation analysis are used to identify the drivers of changes in relation to the observed trends and identifying the tipping points. Environmental Kuznets curves are used to show the relationship between wealth and environmental degradation over the time. This curve hypothesises that at an early stage of development, environmental degradation increases, but beyond some level of economic growth, environmental degradation declines. Further details on the study area and methodology are given in SI.
Reg Environ Change (2016) 16:429–443
Journal