The field and laboratory methods used in this study and discussed as below:
Field Sampling Methods: The Sundarbans Reserved Forest is located at the south west corner of the Ganges River Delta close to the Bay of Bengal, mainly on the sea-shore line, river banks, channels, and small creeks.
Site Selection and Location of the Study Area: The location of these compartments within the Chandpai area, and the location of this area in the wider Sundarbans. Fieldwork was performed in October, 2003 to March, 2004. Locations of sampling points were determined using a Global Positioning System with a precision of 5-10 m. For one typical plot, in compartment 31, the altitude was recorded as 4.4 m above sea level. Among the three compartments, compartment number 26 was selected as an area highly affected by top-dying, where most of the trees were affected severely. The sampling was conducted in a randomized block design, in that a plot was sited within a particular top-dying intensity block, but the precise location of that plot was randomized so as not to collection. Thus in total nine plots were sampled, representing a range of top-dying intensities. Intensive field data collection was made among these nine selected plots. Observations were performed from observation towers during low and high tides, also traversing the forest floor and vegetation on foot, as well as using a speed boat, trawlers, country required to gain access.
Vegetation Recording Methods in the Field: Within each of the nine 20m x 20m plots, each adult tree was assessed for three parameters. The diameter at 1 m height was recorded (in cm) by using a tree diameter- measuring tape or slide calipers depending on girth.
Soil and Water Sampling Methods: As stated above, from the three selected compartments, a total of nine plots of 20m x 20m were selected. Fro m each of these plots, seven soil samples were collected; one from the centre of the plot, four (one each) from all the corners, and two from the middle sides of the plot. This field sampling method followed the W.H.O, U.K, and E.P.A systems of standard laboratory and field sampling principles, rules and regulations.
Questionnaire Survey of Local People: In order to establish the views of local people about the incidence and causes of top-dying, a questionnaire was prepared for asking peoples either individually or in groups. This survey was done among people living or working in the 17 Sub-Districts of Sundarbans, making a distinction between those living within and outside of Sundarbans. They were asked whether they had seen the top-dying disease of Heritiera fomes (Sundri) in Sudarbans for a long time, either through living within the Sundarbans or through visiting Sundarbans for their daily work, for their professional work such as forestry officials, for fishing or for collecting wood as wood cutters, for seasonal honey collection, or other purposes. In this way, 50 questionnaires were filled up through interview, mostly of groups and sometimes of individual people.
Statistical Analysis: Initial statistical analysis of quantitative data, particularly of the elemental concentrations, consisted of calculation of arithmetic means, standard deviations and standard error values for each variable separately. Data on the severity of top-dying for each tree in a plot, which had been recorded as ‘not affected’, ‘mildly affected’, ‘moderately affected’, and ‘highly affected’, were converted into a four-point scale (0-3), so that they could be summed and an average (median) could be determined for each plot, thus producing an index based on ranked data. Comparisons of the strength of relationship between two variables were assessed by correlation: the Pearson’s product-moment correlation coefficient wh ere both variables were fully quantitative or the Spearman’s rank correlation coefficient where the top-dying index was one of the variables. In the case of the Spearman’s coefficient, the probability of the outcome was determined by using the approximation to a t-statistic appropriate to these tests (Sokal and Rohlf, 1981). Occasionally, a Pearson’s correlation coefficient was calculated where top-dying was one of the variables, in order to check on the extent of the difference between the rank and quantitative versions for these data. Data on frequencies of seedlings or saplings in each of the plots and compartments were tested by χ2; contingency table analysis to determine whether there was an association between the selected plot type (severely, moderately or little affected by top-dying) and the three chosen compartments. A similar consideration of the different compartments as comprising one factor, and the plot type as a second, was used to test the pattern of elemental concentration s and other variables by a 2-factor analysis of variance test w ith replication. This allows an assessment of the significance, not only of the two factors separately but also of the interaction term linking the two factors. It should be noted th at the plot type was not a strictly controlled factor, since th e three categories of top-dying intensity were relative to each other within any one compartment and might not have been exactly equivalent between the three categories; interpretation of the results from these tests therefore needs to bear th is in mind. MINITAB Release 14 Statistical Software has been us ed for windows on CD-ROM, 2004 edition for all data analysis, both statistical and graphical, except for those produced automatically by the Excel package attached to the ICPMS.