Description of the study site
Bangladesh is located between 20°34'-26°3' N and 88°01'-92°41' E. It is bordered by the Bay of Bengal on the South and by India on all other sides along with small part of Myanmar in the south-eastern edge. The study is conducted in coastal Khulna region which comprises of three administrative units i.e. Khulna, Jessore and Satkhira district. Study area covers an estimated area of 10,830 sq km) in the South-west corner of Bangladesh. The Khulna region is basically a low lying flat, and fertile deltaic plain, most parts of which are roughly one meter above the mean sea level (MSL). The world famous mangrove forest Sundarban’s is located at the southern edge of this study site. Both calcareous and noncalcareous alluvium soils are found in this region. Tidal influence and salinity intrusion are pronounced in many parts of this region. In fact, strong salinity gradient in soil and water is observed from Jessore (located in the north edge) to Satkhira (located in the south edge). A tropical to subtropical humid monsoonal climate characterizes the entire area. The region experience about 1650 mm rainfall annually with high concentration during the month of June to August. Three distinct seasons such as summer (March–May), rainy (June– October), and winter (November–February) are pronounced in the entire area. The mean annual temperature is 26?C (range: 19–32?C). In some places, temperatures drop to 10?C during the winter and reach 40?C or more during the summer. Cropland agroforestry practices are greatly shaped by many of these climatic factors and non-climatic factors in this region.
Sampling design and data collection
Every district consists of a number of smaller administrative units called sub-district. Twelve subdistricts (4 sub-districts from each district) were selected randomly. In the absence of systematically documented information on cropland agroforestry in each sub-district, cropland agroforestry plots were selected in an unbiased manner following Kabir and Webb (2009). Every day a new local guide was hired to assist in selecting plots of cropland agroforestry. After selecting a cropland agroforestry plot, the guide was requested to stay away from the interview process in order to prevent bias while selecting the next plot of cropland agroforestry. Thus data and information were collected from a total of 313 plots of cropland agroforestry selected randomly during April –June 2013. The questionnaire broadly includes inquiry about the socio-demographic profile of the plot owners, various socio-environmental and ecological attributes of the plots employed for cropland agroforestry, species composition, management practices, and the problem encountered in the cropland agroforestry practices. Various descriptive and inferential statistics were used to analyze the situation.
Determination of crop diversity index
There exist a myriad of methods of computing diversity index. One of the most widely used diversity index is Shannon's index (a measure of biodiversity in an ecological community) which is an ideal one for computing diversity of a particular area but not suitable for comparison purpose. Shannon diversity index allows comparison among the species of two communities only if total sample sizes are equal for both communities. For three districts of Khulna region as the sample sizes are not the same, therefore this index is not used to determine crop diversity. Another widely used diversity index is Simpson's diversity index which takes into account the number of species present, as well as the relative abundance of each species. However, to compute this index the number of species in any area, as well as the number of individuals of each species is required. As this study deals with the varieties of crops species (cereal, non-cereal, spices, vegetables etc.) which could not be counted rather amount of areas designated for each of these crops production could be computed, therefore the main Simpson diversity index could not be used. However, Gini-Simpson index which has similarities with The Gibbs-Martin index of diversity has solved this problem. Because Gini-Simpson index gives the probability those two randomly sampled individuals from the assemblage represent two different species. Therefore, for this research to compute the crop diversity, Gini-Simpson index is used.This research chooses to study the diversity of agricultural crops at the farm/plot level as data is collected for the number of crops cultivated and the area under each crop. From 313 sample plots a total of 150 ha land was included for the computation of crop diversity indexes (CDI) in the study region of Khulna.
Determination of physical properties and chemical indicators of soil
Top soils (up to 10 cm depth) from selected cropland agroforests of each districts was collected using core sampler of 5 cm diameter. Bulk density was determined. Conductivity (EC) of soil was measured. Soil pH was measured. Organic matter of soil was measured. The plant available form of nitrogen in soil was prepared and the plant available form of phosphorus and potassium in soil was prepared . Then the sample extacts were processed to measure nitrogen and phosphorus concentration in sample extracts respectively using UV-Visible Recording Spectrophotometer (U- 2910, HITACHI, Japan). Potassium concentration in sample extracts was measured by Flame photometer (PFP7, Jenway LTD, England).
Statistical analysis
The statistical analysis (multiple responses, χ² test, ANOVA, descriptive statistics) was performed using SPSS (17.0) statistical software.