Data collection and analyses
The research was undertaken between October 2012 and June 2013. A community meeting
was held in the Agrovila Pepital to inform residents about the aims and methods of the study,
and to ask for their participation in the research. Residents of the other villages surveyed were
invited and transportation was made available.
Qualitative methods were used to assess information about livelihoods and local perception
of river degradation, well-being and restoration needs. Semi-structured interviews were
conducted with 79 households (23% of the village families). Interviews were always
preceded by a presentation of the project and the team. The sampling was randomized. The
first house of the village was visited and every fourth thereafter. When a house was empty it
was replaced by the previous house; and if that one was also empty it was replaced with the
next house in the village. Interviewed persons were preferentially old enough to remember
the life and environmental conditions before the relocation process (age averaged 57 years
old, with a standard deviation of 14.5). For this reason, eight questionnaires did not suit the
criteria and were invalidated.
Interviews were conducted on a one-on-one basis and their duration varied between 40
minutes to 2 hours. Interviews consisted of asking relocated residents about four main
subjects: A) their history, family, housing conditions, and perception of well-being; B) their
livelihood and income (families that depended on agriculture and/or had agroforestry home
gardens were asked to freely list the plant species they used); C) their present and past
relation with the river (uses and benefits) and their perception about the situation of the river,
its causes and consequences for human health; and D) their perception about forest
restoration needs and potential approaches. These subjects were presented to the respondents
through a standardized questionnaire that was filled out manually. The responses were
carefully checked and grouped into categories defined afterwards for descriptive statistics
such as percentages and frequencies.
Ethnobotanical information on native forest species and their use was assessed with 19 local
specialists. These local specialists were identified among the respondents as the individuals
that considered themselves and/or were considered by the community to be specialists in
forest species [13]. With them the free-listing technique [14] was used by which they freely
listed the tree species they knew to occur in the Pepital River riparian forest.
The free-list method is an efficient tool to indicate which species belong in the cultural
domain [15]. Smith’s index is a measure of salience that ranges from 0 to 1 and is based on
the highest frequency values and greater coincidence in the position of citation between
informants, thereby promoting ordering of items in the list, which allows the identification of
possible ‘breaks’ in the dataset [16]. A cultural consensus analysis of the data obtained from
the free-list was performed to measure the degree of agreement among informants
(“culturally correct” information) [17]. The frequency and Smith’s index of salience of the
free-list results (agricultural crops, home garden trees species, wild fauna and tree species of
the Pepital River) were calculated by ANTHROPAC 4.9 [15,18], and correlated at > 0.97.
This high correlation between item frequency and Smith’s salience indicates consistency in
the free-list results [19]. Only the species with Smith’s index higher than 0.1 were reported. Field observations and ecological surveys provided complementary information. For the
ecological survey, three plots of 1,000 m2
(50 x 20 m) were established in the most conserved
riparian forests of the Pepital River to evaluate the relative richness and abundance of trees
species (more than 10 individual per hectare). The location of the plots was set after a field
trip carried out alongside the river led by two local specialists. Plots were at least 1 km apart
from each other to guarantee sampling independence. In each plot, all trees >10 cm diameter
at breast height (DBH) were sampled. The common names were given by the local specialists
and samples were brought to the Maranhao State University (UEMA) herbarium to undergo
botanical identification.