The Hierarchical Structure of the Survey¶
GLOW is an organization comprised of various research institutions, community groups, and NGOs, hereafter collectively referred to as local teams. These teams vary in size and capacity, ranging from small local groups to large regional hubs. Regardless of their size or resources, it is imperative that all local teams follow a common hierarchical experimental design. While local logistical constraints (e.g., finite equipment or staffing) will dictate the volume of data collected, the structure of that collection must remain uniform.
Our experimental design employs three nested spatial scales, arranged from smallest to largest: Points, Sites, and Locations.
Points: A point represents the specific field of view surveyed by a single camera. Our cameras focus on an area of approximate 0.25 m^2, which constitutes the smallest indivisible spatial unit in our survey. Points within a single Site must be separated by a minimum of 2–3 meters to ensure independence, and must not be separated by more than 50 meters.
Note: We refrain from calling this a “sample” because, in longitudinal studies (where populations are observed over time), the definition of a sample encompasses both the spatial unit and the specific period of observation.
Sites: A Site is defined as a distinctive mangrove area of no less than 100 m² that is homogeneous in its biological structure and coarse environmental conditions. That is, despite small-scale variations in vegetation or microhabitats, the Site is a unit subject to the same level of human impact (if any) and environmental exposure.
Rules: A Site must contain at least 6 Points (see Canopy Treatment below). Sites can be classified into one of three status levels: Natural, Degraded, or Rehabilitated/Restored. Sites within a Location must be separated by no more than 20 km, but must be distinct enough to avoid overlap with adjacent sites (i.e., the distance between Site A and Site B must be greater than the spread of points within Site A).
Locations: A Location is a broad geographic region within a local team’s area of reach that contains suitable Sites. Distinct Locations should be separated by at least 50 km to ensure regional independence.
Experimental Treatments¶
The design includes two treatments: one orthogonal (independent) and one nested.
Canopy (Orthogonal): This two-level treatment defines the microhabitat setting: under uanopy (forest) vs. upen mudflat (no canopy). This treatment is orthogonal to the Site, meaning every Site must contain both canopy levels. Each level should be represented by at least three Points, resulting in a minimum of 6 Points per Site.
Status of Site (Nested): This treatment defines the condition of the Site: Natural, Degraded, or Restored. Ideally, each Location should include paired comparisons of these levels (e.g., Natural vs. Restored, or Natural vs. Degraded).
Surveys Over Time¶
There are three critical temporal aspects to the survey: when to survey, how often (frequency), and for how long (duration).
We recognize that not all local teams can survey with the same intensity. Furthermore, because teams will join the program at different times, there is no expectation that global surveys will commence synchronously. Decisions regarding timing and intensity must be informed by local knowledge, environmental context, and capacity.
When to Survey?¶
Surveys can commence as soon as the local team has acquired equipment, conducted a pilot study, and verified that camera rigs are operating correctly. However, surveys should be avoided during times of high risk to equipment or personnel, such as cyclone seasons or periods of extreme flooding.
Frequency and Duration Frequency (how often a survey is repeated) and Duration (how long the cameras record during a single survey) are inversely related. We strongly encourage prioritizing duration over frequency. That is, in the context of the GLOW project it is scientifically more valuable to conduct fewer surveys that capture long, continuous periods of data (e.g., weeks) than to conduct many short, fragmented surveys.
The final decision on sampling intensity is conditioned by equipment availability, staffing, and security risks (see Table). In areas where the risk of equipment theft or loss is low, local teams should prioritize deployments that last for several consecutive weeks.
Table : Recommended survey durations based on risk and resources. This table outlines the recommended deployment strategies based on the security of the site and the resources available to the local tea.