Biodiversity survey and estimation for line-transect sampling

Area of Science:

• Ecology
• Biodiversity research
• Statistical modeling

Background:

• Biodiversity surveys often face practical constraints, leading to non-random sampling designs.
• Common biodiversity estimators assume independent sampling, which can be violated in practice.
• Non-independent sampling can compromise the accuracy of biodiversity estimates.

Purpose of the Study:

• To develop a statistical method that accounts for non-independent sampling in biodiversity surveys.
• To improve the accuracy of biodiversity estimation in ecological field studies.
• To provide a solution for analyzing data from sequential sampling methods like line transects.

Main Methods:

• Utilized a first-order Markov chain model to represent sequential sampling.
• Derived a parameter estimator to quantify the degree of non-independence.
• Applied the model to empirical datasets from tropical and subtropical forests and Acacia species distribution.

Main Results:

• The proposed Markov chain model and estimator performed well in empirical tests.
• A significant non-independent sampling effect was detected in subtropical regions using line transects.
• The method demonstrated improved diversity estimation by accounting for sampling dependencies.

Conclusions:

• The Markov chain approach offers a robust solution for biodiversity estimation with non-independent sampling.
• This method is particularly relevant for field surveys employing sequential sampling techniques.
• Accurate biodiversity assessment requires acknowledging and modeling sampling non-independence.