Cr

RCa

b

c

d

e

Rf

a

ARRA

KHISSS

1

oahibrfdpd

b

h1

Ecological Indicators 45 (2014) 219–226

Contents lists available at ScienceDirect

Ecological Indicators

j o ur na l ho me page: www.elsev ier .com/ locate /eco l ind

omparative performance of incidence-based estimators of speciesichness in temperate zone herpetofauna inventories

aluca Ioana Bancila a,b,c, Dan Cogalniceanub, Rodica Plaias ua,b,∗, Marian Tudorb,onstantin Cazacud, Tibor Hartele,f

“Emil Racovita” Institute of Speleology, Romanian Academy, Calea 13 Septembrie 13, 050711 Bucharest, RomaniaUniversity Ovidius Constanta, Faculty of Natural Sciences, Al. Universitatii, corp B, Constanta, RomaniaInstitute of Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190, CH-8057 Zurich, SwitzerlandDepartment of Systems Ecology and Sustainability, University of Bucharest, Spl. Independentei 91-95, 050095 Bucharest, RomaniaSapientia Hungarian University of Transylvania, Faculty of Sciences and Arts, Department of Environmental Studies, Calea Turzii 4, 0264 Cluj-Napoca,omaniaInstitute of Ecology, Leuphana University, 21335 Lüeneburg, Germany

r t i c l e i n f o

rticle history:eceived 5 January 2014eceived in revised form 25 March 2014ccepted 3 April 2014

eywords:erpetofauna

nventoryample unitspecies accumulation curvespecies richness estimators

a b s t r a c t

Species inventories provide the basic information for biodiversity studies. Designing cost-effectivespecies inventories is important in estimating the species richness or the number of additional speciesexpected in a study area. Complete species inventories are difficult to achieve and a variety of estimationmethods are available to counter the underestimation of species richness associated with incompletesampling. A main criterion of selecting an estimation method requires deciding on a sampling unit type.The sampling unit may vary across study and estimating species richness using different sample unitsmay produce different results. We evaluated the ability of incidence-based estimators of species richnessto provide reliable estimates of species richness of temperate herpetofauna (i.e. amphibian and reptile)communities using two different types of sampling units (i.e. numbers of sampling days and the numberof sampling sites, respectively). Our results showed that incidence-based estimators varied in their per-

formance according to the two sampling units used. The incidence-based coverage estimator performedbest when sampling unit was the number of sampling days whereas second-order Jackknife was thebest estimator using sites as sampling unit. We draw attention on the lack of robustness of incidence-based estimators to sampling unit and recommend using species richness estimators in inventories of

on th

herpetofauna depending

. Introduction

The worldwide amphibian and reptile decline is a major issuef concern (e.g., Beebee and Griffiths, 2005; Böhm et al., 2013). Tossist amphibian and reptile conservation much of the researchas focused on developing and implementing inventory and mon-

toring programs. Despite differences in morphology, reproductiveiology and ecological and behavioral traits, amphibians andeptiles are often combined in a single group, called herpetofauna,or inventory and monitoring purpose. A critical issue in the

evelopment and implementation of inventory and monitoringrograms for herpetofauna is selecting the appropriate samplingesign to reduce the uncertainty in estimation the status or trend

∗ Corresponding author at: Emil Racovita Institute of Speleology, Calea 13 Septem-rie 13, 050711 Bucharest, Romania. Tel.: +40 0213188132.

E-mail address: rodica plaiasu@yahoo.com (R. Plaias u).

ttp://dx.doi.org/10.1016/j.ecolind.2014.04.005470-160X/© 2014 Elsevier Ltd. All rights reserved.

e sampling unit.© 2014 Elsevier Ltd. All rights reserved.

of amphibians and reptiles communities. A variety of samplingmethods and units can be employed to inventory and monitor theherpetofauna communities, but due to limited resources availablethere is a critical need for standardized and optimized samplingprotocols. Standardized protocols guarantee the comparabilityof data across studies and communities, whereas optimizationguarantees the efficiency, so that time and resources are notreduced (Lindenmayer and Likens, 2010).

Species richness assessment is the primary step in field stud-ies (Boulinier et al., 1998) and the most commonly used indicatorto describe the diversity on local (e.g. community) and regional(e.g., species pool) level (Colwell and Coddington, 1994; May, 1988).Although species richness is a natural measure of biodiversity, itis an elusive quantity to measure properly (May, 1988). Species

richness is defined as the actual number of species present in agiven area (Brown et al., 2007). A complete census of the numberof species in an extensive area is theoretically possible, but gener-ally not feasible especially in species-rich taxa. Consequently, one

220 R.I. Bancila et al. / Ecological Indicators 45 (2014) 219–226

Table 1Characteristics of the five study areas.

Name Short name Country Latitude (north) Longitude (east) Altitudinal range (m) Area (ha)

1 Macin Mountains National Park Macin Romania 45◦12′ 28◦21′ 7–467 11149.152 Siriu Mountains Siriu Romania 45◦28′ 26◦14′ 548–1663 6230

ania

fiatovstdWmbds

TTi

*

3 Defileul Jiului National Park Jiu Romania

4 Maramures Mountains Natural Park Maramures Romania

5 Western Black Sea Coast Black Sea Coast Bulgaria and Rom

rst assesses the completeness of the species inventory. Speciesccumulation curves (SACs) (Soberón and Llorente, 1993), in whichhe cumulative number of species is plotted against a measuref the sampling effort (e.g., number of samples, number of indi-iduals), are commonly used to evaluate the completeness ofpecies inventory. When SACs reaches a horizontal asymptotehe sampling effort is considered sufficient to capture the speciesiversity in a given area and the species inventory completed.hen the SACs show an increase up trend one uses an estimation

ethod to extrapolate from the data and find the ‘true’ num-

er of species. A large number of estimation methods have beeneveloped (Cogalniceanu, 2007) but incidence-based estimators ofpecies richness are the most widely used.

able 2he observed and the expert extrapolation of total species richness of amphibians and rnventoried but possibly present.

Species Order Study area

Macin

Bombina bombina Anura 1

Bombina variegata Anura 0

Pelobates fuscus Anura 1

Pelobates syriacus Anura 1

Rana arvalis Anura 0

Rana dalmatina Anura 1

Pelophylax kl. esculentus Anura 1

Pelophylax lessonae Anura 1

Pelophylax ridibundus Anura 1

Rana temporaria Anura 0

Bufo bufo Anura 1

Pseudepidalea viridis Anura 1

Hyla arborea Anura 1

Salamandra salamandra Caudata 0

Ichtyosaura alpestris Caudata 0

Triturus cristatus Caudata 0

Triturus dobrogicus Caudata 1

Lissotriton montandoni Caudata 0

Lissotriton v. vulgaris Caudata +

Emys orbicularis Testudines 1

Testudo graeca Testudines 1

Testudo hermanni Testudines 0

Anguis fragilis Sauria 1

Ophisaurus apodus* Sauria 0

Ablepharus kitaibelii Sauria 1

Darevskia praticola Sauria 0

Eremias arguta Sauria 0

Lacerta agilis Sauria 0

Lacerta trilineata Sauria 1

Lacerta viridis Sauria 1

Podarcis taurica Sauria 1

Podarcis muralis Sauria +

Zootoca vivipara Sauria 0

Coronella austriaca Serpentes 1

Elaphe sauromates Serpentes 1

Dolicophis caspius Serpentes 1

Natrix natrix Serpentes 1

Natrix tessellata Serpentes +

Vipera ammodytes Serpentes 1

Vipera berus Serpentes 0

Zamenis longissimus Serpentes 1

Total number of observed species 24

Number of species possibly present 3

Total 27

Known only from the herpetofauna of Bulgaria.

45◦16′ 23◦23′ 234–1690 11,15647◦46′ 24◦33′ 340–1957 150,00043◦24′ 28◦08′ 0–70 12,000

Many studies have evaluated the performance of those estima-tors using indicators such as bias, precision and accuracy underdifferent conditions and sample sizes (Hortal et al., 2006; Merloet al., 2010; Walther and Moore, 2005, for a review). However,these studies used data obtained with similar sampling units. Toour knowledge there is no evaluation of how the incidence-basedestimators perform when different type of sampling units are used.

The aim of this study was to evaluate the performance ofnon-parametric incidence-based estimators of species richness

in herpetofauna inventory in a temperate zone, based on twodifferent sampling units: sampling days and sampling sites. Weevaluated the performance of the estimators using five qualityindicators: bias, precision, accuracy, the smallest sample size

eptiles in each of the five studied areas: 1 indicates presence, 0 absence and + not

Siriu Jiu Maramures Black Sea Coast

0 0 0 11 1 1 00 + 0 +0 0 0 10 0 + 0+ 1 + ++ 0 + 10 0 0 00 1 0 +1 1 1 01 1 1 0+ 1 + 1+ 1 + 11 1 1 01 + 1 01 1 + 00 0 0 0+ 0 1 01 1 + 0+ + + 10 0 0 10 0 0 +1 1 1 00 0 0 10 0 0 +0 1 0 00 0 0 01 1 1 00 0 0 10 1 0 10 0 0 1+ 1 0 11 1 1 0+ + 1 00 0 0 +0 0 0 11 1 1 11 + 0 +0 1 0 11 1 1 0+ 1 0 +13 20 12 159 5 8 822 25 20 23

al Indi

rl

2

2

sbBle2

tss

eatftAs

F(

R.I. Bancila et al. / Ecologic

equired to estimate the observed richness in the total sample andack of erratic behavior in the curve shape.

. Material and methods

.1. Species data

The incidence (presence/absence) data sets used in thistudy were collected during herpetofauna inventories conductedetween 2003–2010 in five protected areas from Romania andulgaria (Table 1). The results of two inventories were already pub-

ished: for Maramures Mountains Natural Park in Cogalniceanut al. (2007) and Western Black Sea Coast (Cogalniceanu et al.,008).

To study how the sampling units affect species richness estima-ors we grouped the data by sampling sites and sampling days. Aite was defined as a location in which observations were made,ituated at least at 100 m from one another.

The true number of species for each studied areas is the expertxtrapolation, i.e. species that were not found during the study butre probably present based on habitat type and species distribu-ion (Cogalniceanu et al., 2013a,b) (Table 2). Different studies have

orecasted that global changes will induce shifts in the range of rep-iles and amphibians that are likely to lose/win (Aragón et al., 2010;raújo et al., 2006; D’Amen et al., 2011; Popescu et al., 2013). At themall spatial scale of the present study, changes in herpetofauna

ig. 1. Species accumulation curves (SACs) plotted against the number of sampling units,

b) Siriu, (c) Jiu, (d) Maramures and (e) Black Sea Coast.

cators 45 (2014) 219–226 221

species richness are predicted by most models. Thus the validity ofa complete species list is very limited in time.

2.2. Data analysis

For each studied area separately on sampling units (days andsites) we calculated six non-parametric incidence-based speciesrichness estimators: incidence-based coverage estimator (ICE),Chao 2 estimator, first-order Jackknife (Jack 1), second-orderJackknife (Jack 2), bootstrap (Boot) and Michaelis–Menten Mean(MMMean), using EstimateS, version 8.20 (Colwell, 2005). Thesample order was randomized 200 times to compute the mean esti-mator and species richness for each accumulation sample level. Wealso constructed species accumulation curves (SACs) to determineif the sampling effort has been sufficient to discover the speciespresent in the studied areas.

To evaluate the performance of the six non-parametricincidence-based species richness estimators we used five indica-tors: (1) the bias (i.e. the difference between the estimates and thetrue number of species), (2) the precision (i.e. the variation of theestimates from all possible samples that can be taken from the com-munity being sampled), (3) the accuracy (a combination of bias and

precision), (4) the smallest sample size required to estimate theobserved richness in the total sample and (5) the lack of erraticbehavior in the curve shape. The bias, precision and accuracy werecalculated following Walther and Moore (2005).

days (solid line) and sites (dashed line) for each of the five studied areas: (a) Macin,

222 R.I. Bancila et al. / Ecological Indi

Tab

le

3Th

e

obse

rved

spec

ies

rich

nes

s

(Sob

s), t

he

exp

ert e

xtra

pol

atio

n

(Eex

p),

the

spec

ies

rich

nes

s

esti

mat

es

and

stan

dar

d

dev

iati

on

for

the

five

stu

die

d

her

pet

ofau

na

com

mu

nit

ies;

N

=

sam

ple

size

; IC

E

=

inci

den

ce-b

ased

cove

rage

esti

mat

or,

Ch

ao

2

=

Ch

ao

2

esti

mat

or, J

ack

1

=

firs

t-or

der

Jack

knif

e,

Jack

2

=

seco

nd

-ord

er

Jack

knif

e,

Boo

t =

boot

stra

p

and

MM

Mea

n

=

Mic

hae

lis–

Men

ten

Mea

n.

Esti

mat

orM

acin

S obs

=

24, E

exp

=

27

Siri

u

S obs

=

13, E

exp

=

22

Jiu

S obs

=

20, E

exp

=

25

Mar

amu

res

S obs

=

12, E

exp

=

20

Bla

ck

Sea

Coa

st

S obs

=

15, E

exp

=

23

Day

s

N

=

124

Site

s

N

=

694

Day

s

N

=

11

Site

s

N

=

148

Day

s

N

=

16

Site

s

N

=

97

Day

s

N

=

18

Site

s

N

=

75

Day

s N

=

17

Site

s

N

=

13

ICE

25.4

4

(0.2

5)

26.5

2

(0.2

8)

14.1

1

(0.8

9)

13.8

1

(0.2

9)

27.4

8

(1.5

1)

22.9

2

(0.6

2)

13.3

6

(0.9

8)

13.8

(0.6

2)

15.1

1

(0.1

8)

15.9

9

(1.2

1)C

hao

2

26.9

2

(4.5

1)

30.0

3

(7.2

7)

13.2

5

(1.0

1)

13.3

6

(0.9

2)

27.4

9

(7.7

5)

21.2

3

(1.8

8)

12.4

(1.1

1)

12.5

1

(1.3

)

15.0

2

(0.1

4)

16.4

3

(2.6

3)Ja

ck

1

26.9

4

(1.7

3)

27.9

9

(1.9

9)

14.7

4

(1.8

2)

14.9

9

(1.3

9)

26.7

1

(2.3

8)

23.9

2

(1.9

6)

13.7

4

(1.2

7)

13.9

4

(2.0

2)

15.2

6

(0.2

6)

17.4

3

(2.5

2)Ja

ck

2

29.8

7

(0.4

2)

31.9

2

(0.5

6)

15.0

6

(1.2

5)

15.0

9

(0.7

8)

31.6

9

(1.4

8)

23.9

7

(1.1

6)

13.9

8

(1.6

5)

14.9

1

(0.9

1)

13.8

1

(1.0

1)

19.0

6

(2.5

0)B

oot

25.1

2

(0.1

9)

25.5

1

(0.1

8)

13.8

3

(0.7

2)

12.9

1(0.

22)

22.5

7

(0.7

7)

22.1

3

(0.3

1)

12.7

9

(0.5

4)

12.8

8

(0.3

2)

15.4

3

(0.1

4)

15.9

8

(1.1

3)M

MM

ean

24.1

8

22.9

7

14.0

4

12.8

5

24.7

3

22.9

8

12.2

4

12.0

8 16

.85

18.4

4

cators 45 (2014) 219–226

For measuring the bias we used the scaled mean error:

Bias = 1An

n∑j=1

(Ej − A

)

where A is the expert extrapolation of the total species richness andEj is the estimated species richness for the jth replicate and n is thenumber of replicate.

For precision we used the complement of the coefficient of vari-ation, the later being the ratio of deviation (SD) and mean (E).

Precision = 1 − SD

E

The accuracy was measured using the scaled square mean erroraccording to formula:

Accuracy = 1 −

⎛⎝ 1

A2n

n∑j=1

(Ej − A

)2

⎞⎠ .

The minimal sample size required to include all the observedspecies richness and the lack of erratic behavior was estimated asin Hellmann and Fowler, 1999. The minimal sample size necessaryto include all the observed species richness was estimated based onthe species richness estimates plotted against the sampling effortfor each estimator and sampling unit. The minimal sample size cor-responds to the linear interpolation of the crossover point wherethe estimators start to overestimate the observed species richness(Hellmann and Fowler, 1999).

3. Results

In all five data sets and for both sampling units species richnessincreased as a function of the sampling effort (Fig. 1), but SACs didnot approach a plateau (i.e. saturation) except for the Black SeaCoast when the number of days was the sampling unit.

The species richness estimators used in this study showed dif-ferences between the estimated number of species and differentresults and performances depending on sampling unit and studiedarea (Tables 3 and 4). All estimators regardless the sampling unitunderestimated (negative bias) the expert extrapolation of the totalnumber of species.

Estimators varied in precision according to sampling units andstudied areas. Estimators tended to be more precise when siteswere used as sampling unit, except for Boot. ICE was the most pre-cise estimator in Macin and Jiu, Boot in Siriu and Maramures andChao 2 in Black Sea Coast when sampling unit was the number ofdays. When sites were used as sampling unit MMMean was themost precise estimator in all communities, except Macin and BlackSea Coast where ICE and Chao 2, respectively, were the most preciseestimators (Table 4).

The estimator ICE was the most accurate estimator when samp-ling unit was the number of days. When sites were used as samplingunit Jack 2 was the most accurate estimator, except for Jiu studyarea (Table 3).

According to the last two performance criteria proposed by Meloand Froehlich (2001) (i.e. smallest sample size and lack of erraticbehavior) Jack 1 had good performance in relation with the con-stancy of minimum sampling size, followed by Boot (Table 5). ICE,MMMean and Jack 2 varied among studied areas and samplingunits. In general ICE showed a tendency of higher sampling effortwhen sampling unit was the number of sites whereas Jack 2 when

sampling unit was the number of days. The Chao 2 had the worstperformance and needed the highest sampling effort.

Jack 1, Jack 2 and Chao 2 lacked the erratic behavior throughstudied areas and sampling units whereas ICE, Boot and MMMean

R.I. Bancila et al. / Ecological Indicators 45 (2014) 219–226 223

Table 4Bias, precision and accuracy of the incidence-based estimators of species richness for the five studied areas using days and sites as sampling units. The minimum values forbias and the maximum values for precision and accuracy are underlined. ICE = incidence-based coverage estimator, Chao 2 = Chao 2 estimator, Jack 1 = first-order Jackknife,Jack 2 = second-order Jackknife, Boot = bootstrap and MMMean = Michaelis–Menten Mean.

Estimator Days Sites

Bias Precision Accuracy Bias Precision Accuracy

MacinICE −0.0054 0.9348 0.9979 −0.0137 0.8905 0.9845Chao 2 −0.0057 0.8887 0.9971 −0.0128 0.8425 0.9803Jack 1 −0.0046 0.8643 0.9973 −0.0107 0.8530 0.9848Jack 2 −0.0027 0.8508 0.9976 −0.0035 0.8459 0.9881Boot −0.0069 0.8487 0.9955 −0.0176 0.8459 0.9736MMMean −0.0064 0.8838 0.9965 −0.0241 0.8594 0.9612SiriuICE −0.3196 0.7856 0.9791 −0.0396 0.8124 0.9835Chao 2 −0.3770 0.7434 0.9716 −0.0452 0.8018 0.9789Jack 1 −0.4115 0.8033 0.9687 −0.0407 0.7950 0.9824Jack 2 −0.4150 0.6499 0.9639 −0.0368 0.8131 0.9855Boot −0.4557 0.8075 0.9625 −0.0464 0.7763 0.9766MMMean −0.4387 0.6660 0.9614 −0.0484 0.8557 0.9766

JiuICE −0.0198 0.8234 0.9715 −0.0138 0.8440 0.9963Chao 2 −0.1282 0.7922 0.9528 −0.0209 0.7993 0.9931Jack 1 −0.1954 0.6856 0.9018 −0.0190 0.7515 0.9925Jack 2 −0.0831 0.6242 0.8818 −0.0114 0.7781 0.9950Boot −0.3171 0.6921 0.8580 −0.0291 0.7242 0.9877MMMean −0.1175 0.7312 0.9334 −0.0174 0.8636 0.9957

MaramuresICE −0.3445 0.8171 0.9870 −0.0401 0.8722 0.9832Chao 2 −0.3822 0.8124 0.9844 −0.0441 0.8535 0.9797Jack 1 −0.3875 0.8286 0.9842 −0.0404 0.8290 0.9825Jack 2 −0.3684 0.7215 0.9839 −0.0365 0.8138 0.9852Boot −0.4336 0.8348 0.9806 −0.0453 0.8207 0.9783MMMean −0.4318 0.7490 0.9798 −0.0427 0.8797 0.9811

Black Sea coastICE −0.2364 0.6877 0.9989 −0.2066 0.7141 0.9912Chao 2 −0.3156 0.8429 0.9989 −0.3133 0.9280 0.9901Jack 1 −0.3411 0.8376 0.9987 −0.3411 0.7692 0.9865

0.9980.9980.998

d(

4

pfttsofmpo

TTiB

Jack 2 −0.3658 0.7285

Boot −0.3717 0.8304

MMMean −0.2814 0.7400

isplayed erratic behavior along all studied areas and sample unitsFig. 2).

. Discussion

Our study showed that the performance of the six non-arametric incidence-based species richness estimators differedor the number of days and sites as sampling units and also amonghe five herpetofauna communities. Also SACs constructed againsthe two types of sampling units generated distinct shapes and con-equently different asymptotes (i.e. surrogates of the true numberf species) for each community. This result suggests that dif-

erent sampling units lead to different species richness patterns

isleading conclusions about the status and/or the trend of her-etofauna communities. Differences in species richness dependingn sampling units may result also in incorrect evaluation of samp-

able 5he minimum sample size measured as percentage ((minimal sample size/total sample sizn the studied herpetofauna communities. ICE = incidence-based coverage estimator, Chaooot = bootstrap and MMMean = Michaelis–Menten Mean.

Estimator Macin Siriu

Days Sites Days Sites Day

ICE 45.16 33.72 27.27 31.76 12.5Chao 2 51.61 50.14 36.36 91.89 43.7Jack 1 25.00 20.61 36.36 47.29 43.7Jack 2 15.32 13.40 36.36 26.35 31.2Boot 50.81 48.27 54.55 79.73 62.5MMMean 53.13 – 18.18 – 12.

3 −0.3199 0.6716 0.98555 −0.4018 0.7666 0.98239 −0.1970 0.6865 0.9906

ling sufficiency and completeness of the species inventories in thestudied area. In our study SAC for the Black Sea Coast area seemsto reach saturation when sampling unit is the number of days, butwhen sites are used as sampling unit the curve maintains an upwardtrend. Although no studies have explicitly compared the day basedSACs with the site based SACs, some authors (e.g. Gotelli andColwell, 2001) suggest that if site-based data are available, theseshould be used because site-based SAC can account for natural lev-els of sample patchiness (i.e. heterogeneity between replicates) inthe data. However, there is not unequivocal support for site basedSACs in the literature. For examples, Ugland et al. (2003) claimedthat the site-based SAC takes into account the variability in num-

ber of species between sites, but does not take into account thefact that within the total area sampled there may be heterogeneitybetween habitat patches. Colwell et al. (2011) noted that samplingis an inevitably continuous exercise and the total species richness

e) × 100)) required by each estimator to approximate the observed species richness 2 = Chao 2 estimator, Jack 1 = first-order Jackknife, Jack 2 = second-order Jackknife,

Jiu Maramures Black Sea Coast

s Sites Days Sites Days Sites

0 16.49 16.66 2.67 35.29 46.155 30.05 44.44 50.67 27.78 23.085 30.93 33.33 37.33 17.65 30.775 17.53 27.78 20.00 70.59 23.080 54.64 55.57 57.33 35.29 46.1550 27.84 33.33 2.67 11.76 15.13

224 R.I. Bancila et al. / Ecological Indicators 45 (2014) 219–226

Fig. 2. The behavior of the species richness estimators: Incidence-based coverage estimator (solid line), Chao 2 estimator (dashed line), first-order Jackknife (dotted line),second-order Jackknife (long dash line), Bootstrap (dot dash line) and MMMean (two dash line) in the five studied areas: (a) and (b) Macin, (c) and (d) Siriu, (e) and (f) Jiu, (g)and (h) Maramures , (i) and (j) Black Sea Coast.

al Indi

imt

laa

sfs(2cuoincn

bpIeeacawelac

(osbtul“srprrF

sirsstseflvraebe

R.I. Bancila et al. / Ecologic

ncreases with the time interval as unrecorded species continuouslyove in the sampled region. Therefore days-based SAC may match

he observed species richness more closely.The sampling unit (e.g. cumulative sampling days per samp-

ing period and cumulative sampling sites per sampling area) had strong effect on the estimates bias, precision and accuracy withinnd among the five herpetofauna communities.

All estimators underestimated the expert extrapolation of totalpecies richness regardless of sample unit or the study area. Theact that estimators underestimated the species richness is con-istent with other studies that tested similar or other estimatorsBrose and Martinez, 2004; Canning-Clode et al., 2008; Cao et al.,004; Petersen and Meier, 2003). However, our results may be theonsequence of the fact that the observed species richness is alsonderestimated in all study areas regardless of sampling unit. Outf the six estimators ICE was the estimator with less estimation biasn all five communities, except Macin when sampling unit was theumber of days. Jack 2 was the most unbiased estimator in all fiveommunities, except Black Sea Cost when sampling unit was theumber of sites.

Regarding the precision and accuracy, ICE and Boot performedest when sampling unit was the number of days, being the mostrecise and accurate estimators. Other previous studies identified

CE as a good estimator of species richness. For example, Chazdont al. (1998) found ICE to outperform all the other investigatedstimators. However, calculation of this estimator requires a vari-ble and usually large number (≥20) of classes of rare species. Thisauses species richness estimates of ICE to vary up to 5% (Walthernd Morand, 1998). Also it is not recommended in inventoriesith low sampling intensity since it has been reported to produce

rratic behavior in the first portion of the sample. When samp-ing unit was the number of sites MMMean was the most precisend ICE was the most accurate estimator in three out of the fiveommunities.

In relation with the two criteria proposed by Melo and Froehlich2001) the Jack 1 estimator performed best. Similar results werebtained by Merlo et al. (2010). The minimal sample size corre-ponds to one of the “stopping criteria” that in general meanseyond a certain sample size the increase of new species inven-oried is low enough to justify not taking a larger size. SAC provedseful in determining the minimal sample size. For minimal samp-

ing assessment using SAC the most usual stopping criteria isseeing the flattening”. Walther and Moore (2005) proposed to stopampling once the variation is less than 5% of the estimated speciesichness whereas Chazdon et al. (1998) proposed to select the sam-le that incorporate a predetermined proportion of the total speciesichness. We estimated the minimal sample size based on crite-ia used in similar studies (Hellmann and Fowler, 1999; Melo androehlich, 2001; Poulin, 1998).

Although our study focused on testing the performance ofpecies richness estimators when different sampling units are usedt is important to draw attention that the performance of the speciesichness estimators could also be affected by the differences in theampling effort used in different study areas (i.e. sampling biasensu Walther and Moore, 2005) and in each sampling unit inhe same study area. Thus studies of how different intensities ofampling effort affects the performance of the species richness inach type of sapling unit should be carefully examined. Despite theact we cannot explicitly distinguish between the effect of samp-ing effort and type of sampling unit, our study design allow toisualize the performance of species richness estimators in a wideange of heterogeneous study areas and sampling effort. A useful

nd good estimator of species richness should combine unbiasedstimates, precision and accuracy in representing the real num-er of species, constancy in the smallest sample size required tostimate the observed number of species and lack of the erratic

cators 45 (2014) 219–226 225

behavior in curve shape among different types of sampling unitand/or sampling effort and/or study areas.

We did not find a single estimator to fulfill all the desirablecriteria but our results allow useful recommendations when usingspecies richness estimators in herpetofauna inventories in temper-ate zone. We recommend the use of the ICE estimator when the typeof sampling unit is the number of days and Jack 2 when sites werechosen as sampling unit. Our results are important for conserva-tion given the urgency with which we must identify herpetofaunacommunities that need to be protected and we recommend testingthe applicability of our results for a broader geographical range.

Acknowledgements

This work was supported by two grants of the RomanianNational Authority for Scientific Research, CNCS-UEFISCDI: POS-DRU/89/1.5/S/58852 (CC) and PN-II-ID-PCE-2011-3-0173 (DC). RIBwas funded by the scientific Exchange Programme Sciex–NMS andTH by Alexander von Humboldt Foundation. We thank the review-ers for their comments that have greatly helped us to furtherimprove our manuscript.

References

Aragón, P., Rodríguez, M.A., Olalla-Tárraga, M.A., Lobo, J.M., 2010. Predicted impact ofclimate change on threatened terrestrial vertebrates in central Spain highlightsdifferences between endotherms and ectotherms. Anim. Conserv. 13, 363–373.

Araújo, M.B., Thuiller, W., Pearson, R.G., 2006. Climate warming and the decline ofamphibians and reptiles in Europe. J. Biogeogr. 33, 1712–1728.

D’Amen, M., Bombi, P., Pearman, P.B., Schmatz, D.R., Zimmermann, N.E., Bologna,M.A., 2011. Will climate change reduce the efficacy of protected areas foramphibian conservation in Italy? Biol. Conserv. 144, 989–997.

Beebee, T.J.C., Griffiths, R.A., 2005. The amphibian decline crisis: a watershed forconservation biology? Biol. Conserv. 125, 271–285.

Böhm, M., Collen, B., Baillie, J.E.M., Bowles, P., Chanson, J., Cox, N., Hammerson, G.,Hoffmann, M., Livingstone, S.R., Ram, M., Rhodin, A.G.J., Stuart, S.N., van Dijk, P.P.,Young, B., Afuang, L.E., Aghasyan, A., Aguayo, A.G., Aguilar, C., Ajtic, R., Akarsu, F.,Alencar, L.R.V., Allison, A., Ananjeva, N., Anderson, S., Andren, C., Ariano-Sanchez,D., Arredondo, J.C., Auliya, M., Austin, C.C., Avci, A., Baker,.P.J., Barreto-Lima, A.F.,Barrio-Amoros, C.L., Basu, D., Bates, M.F., Batistella, A., Bauer, A., Bennett, D.,Böhme, W., Broadley, D., Brown, R., Burgess, J., Captain, A., Carreira, S., Castaneda,M.R., Castro, F., Catenazzi, A., Cedeno-Vazquez, J.R., Chapple, D., Cheylan, M.,Cisneros-Heredia, D.F., Cogalniceanu, D., Cogger, H., Corti, C., Costa, G.C., Couper,P.J., Courtney, T., Crnobrnja-Isailovic, J., Crochet, P.-A., Crother, B., Cruz, F., Daltry,J., Daniels, R.J.R., Das, I., de Silva, A., Diesmos, A.C., Dirksen, L., Doan, T.M., Dodd,K., Doody, J.S., Dorcas, M.E., Duarte de Barros Filho, J., Egan, V.T., El Mouden, E.H.,Embert, D., Espinoza, R.E., Fallabrino, A., Feng, X., Feng, Z.-J., Fitzgerald, L., Flores-Villela, O., Franca, F.G.R., Frost, D., Gadsden, H., Gamble, T., Ganesh, S.R., Garcia,M.A., Garcia-Perez, J.E., Gatus, J., Gaulke, M., Geniez, P., Georges, A., Gerlach, J.,Goldberg, S., Gonzalez, J.-C.T., Gower, D.J., Grant, T., Greenbaum, E., Grieco, C.,Guo, P., Hamilton, A.M., Hare, K., Hedges, S.B., Heideman, N., Hilton-Taylor, C.,Hitchmough, R., Hollingsworth, B., Hutchinson, M., Ineich, I., Iverson, J., Jaksic,F.M., Jenkins, R., Joger, U., Jose, R., Kaska, Y., Kaya, U., Keogh, J.S., Köhler, G.,Kuchling, G., Kumlutas, Y., Kwet, A., La Marca, E., Lamar, W., Lane, A., Lardner, B.,Latta, C., Latta, G., Lau, M., Lavin, P., Lawson, D., LeBreton, M., Lehr, E., Limpus,D., Lipczynski, N., Lobo, A.S., Lopez-Luna, M.A., Luiselli, L., Lukoschek, V., Lund-berg, M., Lymberakis, P., Macey, R., Magnusson, W.E., Mahler, D.L., Malhotra,A., Mariaux, J., Maritz, B., Marques, O.A.V., Marquez, R., Martins, M., Masterson,G., Mateo, J.A., Mathew, R., Mathews, N., Mayer, G., McCranie, J.R., Measey, G.J.,Mendoza-Quijano, F., Menegon, M., Metrailler, S., Milton, D.A., Montgomery, C.,Morato, S.A.A., Mott, T., Munoz-Alonso, A., Murphy, J., Nguyen, T.Q., Nilson, G.,Nogueira, C., Núnez, H., Orlov, N., Ota, H., Ottenwalder, J., Papenfuss, T., Pasach-nik, S., Passos, P., Pauwels, O.S.G., Pérez-Buitrago, N., Pérez-Mellado, V., Pianka,E.R., Pleguezuelos, J., Pollock, C., Ponce-Campos, P., Powell, R., Pupin, F., QuinteroDíaz, G.E., Radder, R., Ramer, J., Rasmussen, A.R.R., Raxworthy, A.R., Reynolds, C.,Richman, R., Rico, N., Riservato, E.L., Rivas, E., Rocha, G., Rödel, P.L.B., RodríguezSchettino, M.-O., Roosenburg, L., Ross, J.P., Sadek, W.M., Sanders, R., Santos-Barrera, K., Schleich, G., Schmidt, H.H., Schmitz, B.R., Sharifi, A., Shea, M., Shi, G.,Shine, H., Sindaco, R., Slimani, R., Somaweera, T., Spawls, R., Stafford, S., Stuebing,P., Sweet, R., Sy, E., Temple, H., Tognelli, M.F., Tolley, K., Tolson, P.J., Tuniyev, B.,Tuniyev, S., Üzüm, N., van Buurt, G., Van Sluys, M., Velasco, A., Vences, M., Vesely,M., Vinke, S., Vinke, T., Vogel, G., Vogrin, M., Vogt, R.C., Wearn, O.R., Werner, Y.L.,Whiting, M.J., Wiewandt, T., Wilkinson, J., Wilson, B., Wren, S., Zamin, T., Zhou,

K., Zug, G., 2013. The conservation status of the world’s reptiles. Biol. Conserv.157, 372–385.

Boulinier, T., Nichols, J.D., Sauer, J.R., Hines, J.E., Pollock, K.H., 1998. Estimatingspecies richness: the importance of heterogeneity in species detectability. Ecol-ogy 79, 1018–1028.

2 al Indi

B

B

C

C

C

CC

C

C

C

C

C

26 R.I. Bancila et al. / Ecologic

rose, U., Martinez, N.D., 2004. Estimating the richness of species with variablemobility. Oikos 105, 292–300.

rown, R.L., Jacobs, L.A., Peet, R.K., 2007. Species richness: small scale. In: Encyclo-pedia of Life Sciences. John Wiley & Sons Ltd, Chichester.

ao, Y., Larsen, D.P., White, D., 2004. Estimating regional species richness using alimited number of survey units. Ecoscience 11, 23–35.

anning-Clode, J., Valdivia, N., Molis, M., Thomason, J.C., Wahl, M., 2008. Estima-tion of regional richness in marine benthic communities: quantifying the error.Limnol. Oceanogr. Methods 6, 580–590.

hazdon, R.L., Colwell, R.K., Denslow, J.S., Guariguata, M.R., 1998. Statistical meth-ods for estimating species richness of woody regeneration in primary andsecondary rain forests of northeastern Costa Rica. In: Dallmeier, F., Comiskey,J.A. (Eds.), Forest Biodiversity Research, Monitoring and Modeling: Concep-tual Background and Old World Case Studies. Parthenon Publishing, Paris, pp.285–309.

ogalniceanu, D., 2007. Biodiversity, 2nd ed. Verlag Kessel, Remagen-Oberwinter.ogalniceanu, D., Bancila, R.I., Samoila, C., Hartel, T., Marian, T., 2007. The herpeto-

fauna of Maramures Mountains Natural Park. In: ENSFI Conference Proceedings,Baia Mare, pp. 300–305.

ogalniceanu, D., Samoila, C., Tudor, M., Skolka, M., 2008. Amphibians and reptilesfrom the Black Sea coast area between Cape Midia and Cape Kaliakra. In: Volumcu lucrarile Conferintei de la Constanta, Editura Ex Ponto, Constanta.

ogalniceanu, D., Rozylowicz, L., Székely, P., Samoila, C., Stanescu, F., Tudor, M.,Székely, D., Iosif, R., 2013a. Diversity and distribution of reptiles in Romania.ZooKeys 341, 49–76.

ogalniceanu, D., Szekely, P., Samoila, C., Iosif, R., Tudor, M., Plaias u, R., Stanescu,F., Rozylowicz, L., 2013b. Diversity and distribution of amphibians in Romania.ZooKeys 296, 35–57.

olwell, R.K., 2005. EstimateS: Statistical Estimation of Species Richness and Shared

Species from Samples. Version 8.2, User’s guide and application published athttp://purl.oclc.org/estimates.

olwell, R.K., Chao, A., Gotelli, N.J., Lin, S-Y., Mao, C.X., Chazdon, R.L., Longino, J.T.,2011. Models and estimators linking individual-based and sample-based rar-efaction, extrapolation and comparison of assemblages. J. Plant Ecol. 1, 3–21.

cators 45 (2014) 219–226

Colwell, R.K., Coddington, J.A., 1994. Estimating terrestrial biodiversity throughextrapolation. Philos. Trans. R. Soc. London, Ser. B 345, 101–118.

Gotelli, N.J., Colwell, R.K., 2001. Quantifying biodiversity: procedures and pitfalls inthe measurement and comparison of species richness. Ecol. Lett. 4, 379–391.

Hellmann, J., Fowler, G., 1999. Bias, precision, and accuracy of four measures ofspecies richness. Ecol. Appl. 9, 824–834.

Hortal, J., Borges, P.A.V., Gaspar, C., 2006. Evaluating the performance of speciesrichness estimators: sensitivity to sample grain size. J. Anim. Ecol. 75, 274–287.

Lindenmayer, D.B., Likens, G.E., 2010. Effective Ecological Monitoring. CSIRO Pub-lishing, Melbourne, Australia.

May, R.M., 1988. How many species on earth? Science 241, 1441–1449.Melo, A.S., Froehlich, C.G., 2001. Evaluation of methods for estimating macroinver-

tebrate species richness using individual stones in tropical streams. FreshwaterBiol. 46, 711–721.

Merlo, M.J., Parietti, M., Etchegoin, J.A., 2010. Evaluation of species richness esti-mators in studies of diversity involving two larval digenean communitiesparasitizing snail hosts. Parasitol. Res. 107, 1093–1102.

Petersen, F.T., Meier, R., 2003. Testing species-richness estimation methods onsingle-sample collection data using the Danish Diptera. Biodivers. Conserv. 12,667–686.

Popescu, V.D., Rozylowicz, L., Cogalniceanu, D., Niculae, I.M., Cucu, A.L., 2013. Movinginto protected areas? Setting conservation priorities for Romanian reptiles andamphibians at risk from climate change. PLoS One 8 (11), e79330.

Poulin, R., 1998. Comparison of three estimators of species richness in parasitecomponent communities. J. Parasitol. 84, 485–490.

Soberón, J., Llorente, J., 1993. The use of species accumulation functions for theprediction of species richness. Conserv. Biol. 7, 480–488.

Ugland, K.I., Gray, J.S., Ellingsen, K.E., 2003. The species-accumulation curve andestimation of species richness. J. Anim. Ecol. 72, 888–897.

Walther, B.A., Morand, S., 1998. Comparative performance of species richness esti-mation methods. Parasitology 116, 395–405.

Walther, B.A., Moore, J.L., 2005. The concepts of bias, precision and accuracy, andtheir use intesting the performances of species richness estimators, with a lit-erature review of estimator performance. Ecography 28, 815–829.