Costs and benefits of changing sampling methods in long-term monitoring of marine protected areas
To confirm their effectiveness, long-term monitoring of marine protected areas (MPAs) is critical. Such longitudinal sampling can, however, pose problems of methodology, particularly as more preferred tools may evolve over time. My research sought to understand how two commonly used methods, line intercept transect (LIT) and photoquadrat (PQ), which were used sequentially in MPA monitoring, estimated percent coral cover in a series of MPAs in the central Philippines.
It can be difficult to reconcile data from LIT and PQ because their different parameters and sample sizes can generate contradicting results. I specifically addressed the issues of (i) small sample sizes of LIT data because of a challenge in maintaining fixed transects and (ii) possible spatial autocorrelation of PQ data owing to a decision to place transects close together. The data also faced other problems, including numerous missing data points and inconsistent data labeling.
Analyzing three years of data when both LIT and PQ were applied, I found that LIT estimated higher mean percent coral cover than PQ. However, subsampling LIT data – and thus increasing sample sizes – improved comparison of estimates by these two methods. This finding demonstrated that PQ data, with larger sample sizes from wider spatial ranges, detected significant temporal patterns of coral cover whereas LIT data did not. This result confirmed that more observations and a wider spatial extent of sampling increase the power to detect statistical significance. I also found that PQ data did not exhibit significant spatial autocorrelation, eliminating one of concerns in analyses.
Two other aspects of monitoring also stand out in their importance for better assessment of coral cover. The first is the need for frequent communication between field biologists and researchers / analysts to assure the proper execution of sampling procedures and reporting of any changes to field methods. The second is the vital need to minimize errors by ensuring the accuracy of data encoding and use of standardized labeling systems over time. The quality of the data determines the accuracy of coral cover estimates, especially when different methods are used in MPA monitoring.
Fujii, I. 2018. Costs and benefits of changing sampling methods in long-term monitoring of marine protected areas. Master of Science dissertation, University of British Columbia, Canada.