416 S. P. Woodruff et al.
TABLE 1 A comparison of costs, and bias and precision (CV and RMSE), between aerial sighting and non-invasive genetic sampling capture–recapture methods.
Sampling method Aerial
True abundance No. of samples1 Unknown4
Genetic capture–recapture 200 200
150 300
Total cost2 (USD) CV3 (%) 10,000
21 5,757 10,075
included in genetic capture–recapture cost. 3CV for aerial methods is based on sightability. Capwire is single session and MARK is two sessions. 42014 population estimate from aerial surveys was 202, 95%CI 171–334.
CV was ,10%in 100%of MARK simulations and 84%of capwire simulations (Supplementary Table 1). Two capwire simulations hadCV.20%, both of which had relatively low sampling rates of #0.75 samples/individual/session. The CI was 0–106% of the true abundance, and overall
MARK had narrower confidence intervals than capwire. The CI in capwire was 0–0.73 probability coverage. The highest probability of coverage came from simulations with the lowest ratio of samples per individual (e.g. 0.5 sam- ples/individual) and coverage probability decreased as the ratio increased (Fig. 1, Supplementary Table 1). Probability of CI coverage in MARK was #0.56. Capture probabilities inMARKwere high (0.29–1.00) and mean capture probabil- ities for males were higher than those of females. High cap- ture probabilities led to extremely precise estimates, which often missed the true abundance by 1 or 2 individuals (i.e. true abundance = 150,estimate = 149,CI 149–149). However, nearly all (99%, n = 82) MARK estimates were within 10%of the true abundance (e.g. 180 or 220 for true abundance of 200).On the contrary, a ratio of three samples per individual was required to achieve #10%biasin capwire. Overall, RMSE decreased with increasing sample size
(Fig. 1, Supplementary Table 1). When comparing total number of samples between the estimators (e.g. 200 total samples in one session for capwire or over two sessions in MARK), RMSE was consistently lower in MARK compared to capwire. Additionally, with larger population size, the gap was larger between RMSE values in MARK and capwire RMSE values with comparable sample sizes. For example, with a population of 300, there were 600 total samples in one session for capwire but spread over two sessions in MARK, and RMSE values indicated better overall perfor- mance with MARK (0.249)than capwire (5.31).
Cost comparison
The cost of the aerial survey isUSD 10,000 annually (i.e. half of the cost of the biennial count; USFWS, 2015). To obtain a CV equal to the CV from the aerial estimate (21%), genetic capture–recapture simulations (population size = 200)
indicated 0.75 samples/individual (confirmed consensus genotypes)were needed in just a single sampling session (cap- wire) at a cost of USD 5,757 (Table 1)withRMSE =
6.48.For the same cost, MARK (two sessions) substantially outper- formed capwire,withCV= 8% and RMSE =
1.45.For the same cost as the aerial survey, estimates from genetic cap- ture–recapture methods (MARK, two sessions: CV ,0.5; RMSE ,2.2) are markedly more robust. At this cost preci- sion is also improved over aerial estimates with capwire (CV ,0.13) although RMSE is much higher (,8.0)than when
usingMARK.Considering only the costs, aerial surveys become more cost effective at a population of c. 350 (cost per sample =USD 28.78, Supplementary Table 2;USD 10,000/ 28.78 = 347).
Discussion
Using an abundance estimator that provides the most accu- rate and precise estimate is particularly important when managing threatened species. A highly precise overestimate of the population would give the impression that the species is more abundant, and could result in the implementation of improper management actions (Noss et al., 2012). Our study illustrates the need to use the appropriate sampling design and estimator in capture–recapture studies, and the inher- ent trade-off between the accuracy and precision of an abun- dance estimate and the effort and cost of monitoring. In our simulations the multi-session MARK model performed bet- ter than the single-session capwire estimator, indicating a clear optimal estimation method. As hypothesized, given the very high precision (CV c. 2%) in our previous genetic capture–recapture study (Woodruff et al., 2016b), substan- tial time and money can be saved by reducing sampling effort (e.g. fewer sessions, fewer samples) with little com- promise in precision. Moreover, the genetic capture– recapture method has the potential to provide additional information on population genetic metrics and survival of different sex and age classes (Woodruff et al., 2016b). Simulations can be used to design an efficient sampling scheme, but they are limited by the inability to exactly
Oryx, 2020, 54(3), 412–420 © 2018 Fauna & Flora International doi:10.1017/S003060531800011X
capwire MARK capwire MARK 21 13
8 ,5 6.48 1.45 ,8.00 ,2.20
1Number of consensus genotypes and represents 75% of the number of samples collected to account for failed samples because of DNA degradation. 2Cost of aerial method includes cost of flight time and pilot but not associated salaries for personnel (USFWS, 2015). See Supplementary Table 2 for what is
RMSE3
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100 |
Page 101 |
Page 102 |
Page 103 |
Page 104 |
Page 105 |
Page 106 |
Page 107 |
Page 108 |
Page 109 |
Page 110 |
Page 111 |
Page 112 |
Page 113 |
Page 114 |
Page 115 |
Page 116 |
Page 117 |
Page 118 |
Page 119 |
Page 120 |
Page 121 |
Page 122 |
Page 123 |
Page 124 |
Page 125 |
Page 126 |
Page 127 |
Page 128 |
Page 129 |
Page 130 |
Page 131 |
Page 132 |
Page 133 |
Page 134 |
Page 135 |
Page 136 |
Page 137 |
Page 138 |
Page 139 |
Page 140 |
Page 141 |
Page 142 |
Page 143 |
Page 144 |
Page 145 |
Page 146 |
Page 147 |
Page 148
