A visual assessment technique for estimating seagrass standing crop

ABSTRACT

1. Ground-truthing techniques for measuring seagrass standing crop need to be simple, precise, non-destructive and quick. Mellors (1991) designed a visual assessment technique for estimating above-ground seagrass biomass. This paper builds on Mellors’ work and presents a modified method for estimating standing crop which uses a six-point ordinal scale (excluding zero).

2. Calibration is made before and after the survey period to encompass changes in observer performance. The method has been calibrated for the Caribbean seagrasses Thalassia testudinum Banks ex König and Syringodium filiforme Kützing.

   A strong linear relationship existed between measured standing crop (square root transformed) and standing crop categories (r² = 0.94, p<0.0001, n=103). Although the overlap between some standing crop categories can reach 8-13% of the total range of standing crop, the division of biomass into a six point scale was tested using single factor ANOVA with Student-Newman-Keuls test for multiple comparisons.

   The scale is not dependent on the combination of seagrass species and should be appropriate wherever the maximum above-ground biomass of seagrass reaches 200-300 g/m². Significant changes in calibration can occur over a six week period. Re-calibration of the scale is recommended if the survey period extends across seasons or if surveys are conducted infrequently. If the survey requires less than 24 samples it is more efficient to adopt conventional quadrat harvest to measure standing crop. However, the visual assessment method conveys considerable savings in time for larger surveys (13-fold for 500 samples).

3. The standing crop of seagrass beds, assessed on an ordinal scale, can be compared using the non-parametric Mood median test. The power of this test was calculated for a range of sample sizes and alternative hypotheses using Monte Carlo simulation. The statistical power of these tests was not reduced by sampling errors provided that the minimum detectable difference between seagrass beds was > or = 2 biomass categories and a minimal sample size of 6 quadrats was used. Power is considerably reduced if a difference of 1 biomass category is desired. As such, the decimal recording scale advocated by Mellors (1991) is considered excessively precise and inappropriate for the seagrass beds sampled in the present study.

   The statistical methods and data presented in this study will allow sample size / power relationships to be estimated for any comparable study and therefore negate the need to conduct pilot surveys

   .