Inferring versus measuring rates of recovery in no-take marine reserves

Can rates of biomass recovery of fished species be inferred reliably from once-only spatial comparisons of no-take marine reserves of different ages and fished areas? We used underwater visual census at 15 no-take marine reserves in the Philippines to both infer and measure such rates. We made a single estimate of the biomass of large predatory fishes (Serranidae, Lutjanidae, Lethrinidae) targeted heavily by fisheries in each of 13 well protected no-take reserves (age range 0.5 to 13 yr), and in nearby nonreserve (fished) sites. We also measured rates of biomass buildup of these fish regularly for 18 yr (1983 to 2001) in 2 no-take reserves (Sumilon, Apo) and nonreserve sites. The duration of protection required to detect significantly higher reserve biomass was similar, but lower for temporal monitoring (3 to 4 yr) than for spatial comparisons (6 yr). The reserve:nonreserve biomass ratios at maximum duration of reserve protection were similar for inferred (9.0) and measured (6.3 to 9.8) estimates. Thus, results of long-term monitoring of 2 reserves may have regional generality. The inferred rate of change of a reserve effect index (log 10 [Reserve biomass + 1 / Nonreserve biomass + 11) with duration of protection did not differ significantly from the measured rate at Sumilon, but was higher than that measured at Apo. A habitat complexity index did not affect estimates of 'reserve effects' significantly in this study, and reserve protection was generally effective. Thus, using similar methods of reserve protection and census on the same target group in similar areas, one can make useful inferences about rates of recovery in no-take marine reserves.