Comparison o f density estimators in coastal populations using a
modelling approach
V oigt Maria123, Hichem G herraz', Uwe G riiters4, Nico Koedam2 and Farid Dahdouh-Guebas12
1 Laboratory o f Systems Ecology and Resource Management, Université Libre de Bruxelles - ULB,
Av. F.D. Roosevelt 50, CPI 2 6 4 /1 ,1 0 5 0 Brussels, Belgium
E-mail: mvoiqt@ vub.ac.be
2 Laboratory o f Plant Biology and Nature Management, Vrije U niversiteit Brussel - VUB, Pleinlaan 2,
1050 Brussels, Belgium
3 Ecology Center, Kiel University, Olshausenstraße 75, House I, D -2 4 1 1 8 Kiel, Germany
4 Institute o f Forest Biometrics and System Analysis, Technical U niversity o f Dresden - TUD, Pienner
Straße 8, D -01737 Tharandt, Germany
The earth is facing a b io dive rsity crisis and coastal ecosystems like mangroves, coral reefs and
seagrass beds are severely declining in spatial e xte n t and com position (Duke et a i , 2007; Stone,
2007; Cressey, 2009). In th is light, the assessment o f these com m unities and th e ir bio dive rsity
becomes increasingly im p orta nt. Reliable estim ates o f population density and related parameters
fo rm an integral part o f such studies, because th ey allow the detection o f changes in species
com position and consequently the conservation o f endangered ecosystems.
D ifferent m ethods have been used to obtain population density estim ates o f non-m otile organism s.
They are based e ither on counting the individuals in an area (plot-based) or on determ ining the
distance between them (plot-less) (Krebs, 1999). While plot-based m ethods are robust and provide
good estim ates when applied to populations w ith d iffe re n t patterns, th e y are labour intensive. Plot­
less estim ators, on the o the r hand, are less tim e consum ing, but often heavily biased (Magnussen et
a i , 2012). Many researchers continue to use plot-less m ethods, even th ough very high errors have
been reported (Hijbeek et a i , 2013). It is thus tim e ly to device to ols to help in the decision about
the o ptim al density estim ation m ethods fo r d iffe re n t study systems.
A m odel was w ritte n in the individual-based m odelling language NetLogo (Khan M.N.I., Gherraz H.,
G riiters U., Dahdouh-Guebas F., unpublished results). It allows the user to create a virtua l
population w ith d iffe re n t co nfig uratio n possibilities fo r p o in t patterns, diam eter d is trib u tio n ,
species com position and zonation. It includes a set o f estim ation m ethods, and the option to
choose between sam pling designs and sizes. The m odel calculates the true density, area and
frequency and a d d itio n a lly provides th e ir estim ation as obtained by the d iffe re n t m ethods. It was
designed considering especially coastal ecosystems w ith th e ir d is tin c t zonation, co m p le x ity and
d iffic u lty o f access. A fte r a p ilo t study in the fie ld , the user can rebuild the ecosystem in the model
and use it to fin d the best estim ation m ethod in term s o f accuracy, precision and efficiency fo r the
respective study area.
In the on-going project, we w ill fu rth e r develop the m odel (1) to include an option to e ither calculate
the sample size based on a least accepted e rro r or calculate the e rro r based on a predefined sample
size, and (2) to calculate the efficiency o f each m ethod. The sam pling e ffo rt, which is usually
assessed based o nly on com m on sense, w ill be estim ated as a fu n ctio n o f the num ber o f individual
m easurem ents executed and the distance covered between them . The upgraded m odel w ill be
rig oro usly tested fo r a num ber o f scenarios com prising d iffe re n t co nfig uratio ns o f the input
parameters. The outcom e o f th is study w ill be a catalogue o f indications fo r density estim ation
m ethods in populations w ith d iffe re n t patterns, considering especially the p articularities o f coastal
ecosystems.
References
Cressey D. 2009. Vital marine habitat under threat. Nature.
Duke N.C., J.-O. Meynecke, S. D ittm ann, A.M. Ellison, K. Anger, U. Berger, S. Cannicci, K. Diele, K.C.
Ewel, C.D. Field, N. Koedam, S.Y. Lee, C. Marchand, I. Nordhaus and F. Dahdouh-Guebas.
2007. A w orld w ith o u t mangroves? Science 317:41 -42.
Hijbeek R., N. Koedam, M .N.I. Khan, J.G. Kairo, J. Schoukens and F. Dahdouh-Guebas. 201 3. An
evaluation o f plotless sam pling using vegetation sim ulations and field data fro m a
mangrove forest. PLoS ONE 8, e 6 7 2 0 1 .
Krebs C.J. 1999. Ecological m ethodology, B enjam in/C um m ings Menlo Park, C alifornia.
Magnussen S., L. Fehrman and W.J. Platt. 201 2. An adaptive com posite density estim a tor fo r k-tree
sam pling. European Journal o f Forest Research 131:307-320.
Stone R. 2007. A w orld w ith o u t corals? Science 316:678-681.
-
127
-