Table 5. Maximum diversity (species richness) estimators applied to the NE Madagascar surveys, based on random samples of 6, 7, 8, 10, 12, 15, 20, 25, 30 and 36 samples*.
Samp | Sobs | Chao | Jacknife | Boot- | UGE | MM | Smax | ||
les | 1 | 2 | 1 | 2 | strap | ||||
6 | 212 | 269.6 | 269.6 | 267.8 | 291.7 | 238.7 | 212 | 211.8 | 286.3 |
7 | 210 | 249.6 | 249.6 | 262.3 | 277.9 | 236.0 | 210 | 210.2 | 286.4 |
8 | 221 | 253.7 | 253.7 | 267.4 | 279.5 | 244.6 | 221 | 221.2 | 282.5 |
10 | 235 | 291.7 | 291.7 | 291.7 | 317.2 | 261.9 | 235 | 231.3 | 285.9 |
12 | 236 | 264.9 | 264.9 | 279.1 | 289.9 | 257.8 | 236 | 234.0 | 279.1 |
15 | 242 | 265.1 | 265.1 | 282.1 | 289.4 | 263.2 | 242 | 243.8 | 298.3 |
20 | 252 | 281.4 | 281.4 | 295.7 | 306.4 | 274.3 | 252 | 248.8 | 290.6 |
25 | 257 | 269.9 | 269.9 | 285.3 | 291.4 | 273.0 | 257 | 255.9 | 287.6 |
30 | 262 | 274.0 | 274.0 | 285.7 | 290.8 | 275.5 | 262 | 263.0 | 292.9 |
36 | 269 | 285.9 | 285.9 | 294.3 | 300.5 | 282.6 | 269 | 268.0 | 292.7 |
max | 269.0 | 291.7 | 291.7 | 295.7 | 317.2 | 282.6 | 269.0 | 268.0 | 298.3 |
m | 239.6 | 270.6 | 270.6 | 281.1 | 293.5 | 260.8 | 239.6 | 238.8 | 288.2 |
sd | 20.7 | 13.3 | 13.3 | 11.8 | 11.8 | 16.3 | 20.7 | 20.5 | 5.6 |
CV | 0.086 | 0.049 | 0.049 | 0.042 | 0.040 | 0.063 | 0.086 | 0.086 | 0.019 |
range | 59.0 | 42.1 | 42.1 | 33.4 | 39.3 | 46.6 | 59.0 | 57.8 | 19.2 |
Sobs - Curve of observed species counts; Chao 1 - Chao's estimator based on number of rare species; Chao 2 - Chao's estimator using just presence-absence data; Jacknife 1 - Jacknife estimator based on species that only occur in one sample; Jacknife 2 - Second order jacknife estimator; Bootstrap - Bootstrap estimator based on proportion of quadrats containing each species; UGE - Calculated species accumulation curve; Michaelis-Menten - MM are points on the curve fitted to Sobs, with curve parameter Smax = asymptote at infinity [39], [88]–[92].
Notes:
The only condition for including a sample in these subsamples was that for cases of <15 samples, only full-diversity samples were selected by excluding minor samples in which <40 coral species were recorded overall (see methods for cases where sampling of species was constrained by other objectives during a dive, so records were taken only of new species seen in that dive).