REVIEW ARTICLE Metapopulation and its applications in conservation biology Koson Yuttham1, Mullica Jaroensutasinee2, and Krisanadej Jaroensutasinee3 Abstract Yuttham, K., Jaroensutasinee, M. and Jaroensutasinee, K. Metapopulation and its applications in conservation biology Songklanakarin J. Sci. Technol., 2003, 25(3) : 395-409 Metapopulation, originally pioneered by Levins, refers to a series or set of local populations, each existing on a patch of suitable habitat that is separated from other occupied patches by unsuitable terrain. Each population has its own relatively independent dynamics. The long-term persistence and stability of the metapopulation arise from a balance between population extinction and recolonisation. Although metapopulation concepts are not commonly fit to real world conditions and there are still questions about the generality of metapopulation dynamics and the applicability of metapopulation models to specific prob- lems in conservation, they have contributed significant insights into conservation and have inspired field studies focused on collecting key data on demography and movement. Moreover, metapopulation concepts have had positive effects in conservation research, especially, as interest in metapopulation dynamics has prompted renewed research concerning dispersal capacities that is extremely valuable in understanding population structure. Therefore, metapopulation concepts replace the island biogeography concepts that have been used in the past for decades. Key words : metapopulation, ecogogical modeling, conservation, minimum viable metapopulation 1Master student in Computational Science, 2Ph.D. (Ecology, Evolution and Marine Biology), Asst. Prof., 3Ph.D. (Computational Physics), Asst. Prof., Institute of Science, Walailak University, Tha Sala, Nakhon Si Thammarat 80160 Thailand. Corresponding e-mail: [email protected] Received, 21 February 2003 Accepted, 4 March 2003 Songklanakarin J. Sci. Technol. Metapopulation & Conservation Biology Vol. 25 No. 3 May-Jun. 2003 396 Yuttham, K., et al. ∫∑§—¥¬àÕ ‚°»≈ ¬ÿ∑∏√√¡ ¡—≈≈‘°“ ‡®√‘≠ ÿ∏“ ‘π’ ·≈– °ƒ…≥–‡¥™ ‡®√‘≠ ÿ∏“ ‘π’ ª√–™“°√·∫∫‡¡µâ“·≈–°“√𔉪„™â„π¥â“π™’««‘∑¬“°“√Õπÿ√—°…å «. ߢ≈“π§√‘π∑√å «∑∑. 2546 25(3) : 395-409 ª√–™“°√·∫∫‡¡µâ“‰¥â¡’°“√‡√‘Ë¡µâπ»÷°…“‚¥¬ Levins À¡“¬∂÷ß Õπÿ°√¡À√◊Õ‡´Á∑¢Õߪ√–™“°√∑’ËÕ“»—¬Õ¬Ÿà°—𠇪ìπ°≈ÿà¡ ª√–™“°√°≈ÿà¡π’Ë·¬°ÕÕ°®“°ª√–™“°√°≈ÿà¡Õ◊ËπÊ ‚¥¬ ¿“ææ◊Èπ∑’Ë∑’ˉ¡à‡À¡“– ¡ ª√–™“°√·µà≈–°≈ÿà¡¡’ æ≈«—µ√¢Õß¡—π‡ÕßÕ¬à“ßÕ‘ √–‰¡à¢÷Èπ°—∫ª√–™“°√Õ◊Ëπ ª√–™“°√·∫∫‡¡µâ“®–Õ¬Ÿà‰¥â·≈–¡’§«“¡‡ ∂’¬√ ŸßÀ√◊Õ‰¡à¢÷ÈπÕ¬Ÿà °—∫§«“¡ ¡¥ÿ≈√–À«à“ß°“√ Ÿ≠æ—π∏ÿå·≈–°≈—∫¡“¬÷¥‡ªìπÕ“≥“π‘§¡¢Õߪ√–™“°√‚¥¬¡’°“√Õæ¬æ‡ªìπµ—«°≈“ß„π°“√∑” Àπâ“∑’Ëπ’È ∂÷ß·¡â«à“·π«§‘¥¢Õߪ√–™“°√·∫∫‡¡µâ“‰¡à§àÕ¬®–‡À¡“– ¡°—∫ ∂“π°“√≥å®√‘ß·≈–¬—ߧߡ’§”∂“¡‡°’ˬ«°—∫ æ≈«—µ√¢Õߪ√–™“°√·∫∫‡¡µâ“·≈–°“√®–π”‡Õ“º≈¢Õß·∫∫®”≈Õß∑“ߧ≥‘µ»“ µ√剪„™â„π°“√·°âªí≠À“∑“ߥâ“π°“√ Õπÿ√—°…剥âÀ√◊Õ ·µàª√–™“°√·∫∫‡¡µâ“¬—ߧߡ’§«“¡ ”§—≠¡“°„π¥â“π°“√Õπÿ√—°…å‡æ√“– “¡“√∂∑”„À⇢Ⓞ®·≈–‡ªìπ ·√ß®Ÿß„®„π°“√»÷°…“¿“§ π“¡ ∑”„Àâ‚ø°— °“√‡°Á∫¢âÕ¡Ÿ≈∑’Ë ”§—≠‡°’ˬ«°—∫°“√»÷°…“¥â“πª√–™“°√ ·≈–°“√Õæ¬æ ¢Õߪ√–™“°√ πÕ°®“°π’È·π«§‘¥¢Õߪ√–™“°√·∫∫‡¡µâ“¬—ß¡’º≈‡™‘ß∫«°°—∫°“√∑”«‘®—¬¥â“π°“√Õπÿ√—°…å ‚¥¬‡©æ“– Õ¬à“߬‘Ëßæ≈«—µ√¢Õߪ√–™“°√·∫∫‡¡µâ“∑’ˇ°’ˬ«°—∫§«“¡ “¡“√∂„π°“√°√–®“¬∑”„Àâ “¡“√∂‡¢â“„®‡°’ˬ«°—∫‚§√ß √â“ß ¢Õߪ√–™“°√¡“°¢÷Èπ ¥—ßπ—Èπ·π«§‘¥ª√–™“°√·∫∫‡¡µâ“®–‡¢â“¡“∑¥·∑π·π«§‘¥°“√»÷°…“¥â“πª√–™“°√‡°“– (island biogeography) ∑’ˇ§¬„™â°—π¡“À≈“¬ ‘∫ªï ”π—°«‘™“«‘∑¬“»“ µ√å ¡À“«‘∑¬“≈—¬«≈—¬≈—°…≥å Õ”‡¿Õ∑à“»“≈“ ®—ßÀ«—¥π§√»√’∏√√¡√“™ 80160 The term metapopulation was first intro- What is a model? A model is a representa- duced by Levins (1969). Studies using meta- tion of a particular thing, idea, or condition (Jackson population generally subdivide the general popu- et al., 2000). Models can be very simple, such as lation into a series of local populations with a a logistic growth model for a single species, or balance between extinctions and recolonisations extremely complex, such as an individual based of local populations that facilitates long-term per- model. Model can be categorised into two classes: sistence of the metapopulation. The key process is qualitative (expressed in words or diagrams) and the interpatch connection functions by migration quantitative (expressed in mathematical equations (Figure 1) (Hanski and Gilpin, 1997). Metapopu- or computer languages) (Wu, 1993). Many eco- lation models have been widely used in the logical theories today are represented in largely biological field including in population ecology, mathematical terms because mathematics provides conservation biology, and pest control (for details the most precise language to describe complex see Wu, 1993; Harrison, 1994; Hanski and Gilpin, ecological systems and is also an ideal tool for 1997; Takagi, 1999; Fagan et al., 2002). prediction in ecological systems (Tilman et al., Modelling has become an important tool in 1994; Jansen, 1995; Gyllenberg and Hanski, 1997; population and conservation biology as evidenced Hanski and Ovaskainen, 2000; Keymer et al., by many books that provide excellent advice on 2000; Casagrandi and Gatto, 2002). However, model construction, that have recently published mathematical formulations have limitation in that by such authors as: Cappuccino and Price, 1995; they usually force ecologists to make clear and Haefner, 1996; Hanski and Gilpin, 1997; Hilborn unambiguous assumptions. There are many models and Mangel, 1997; Roughgarden, 1998; Shugart, used in population ecology including metapopula- (1998). tion models. Metapopulation models are useful Songklanakarin J. Sci. Technol. Metapopulation & Conservation Biology Vol. 25 No. 3 May-Jun. 2003 397 Yuttham, K., et al. Figure 1. The ideal metapopulation is a set of local populations in which the interpatch con- nection functions by migration facilitate long-term persistence with the balance between colonisation and extinction. The empty habitat patches (in white) are susceptible to colonisation from nearby local population. Local populations whose persistence depends on immigration from nearby local populations (sink popula- tions) are shown in black, while the sources of immigrants (source populations) are shown in grey (Marquet, 2002). and popular models that fall into the quantitative In this review article, we will briefly des- type and are always represented as analytical or cribe the metapopulation concepts and how meta- simulation models. population concepts are applied in conservation There are three types of modelling (for excellent detail see Hanski and Gilpin, 1997). approaches used in metapopulation studies Finally, we will give some ideas about the use assuming many habitat patches and local popula- and misuse of metapopulation studies, especially tion (Hanski and Gilpin, 1997): (1) spatially implicit concerning rules for both theoreticians and con- approaches, often based on a critical simplification servationists which have been argued by Doak and of what at first appears as a hopelessly complex Mills (1994). problem, in which the habitat patches and local populations are discrete (and are generally assumed Mathematical Modelling to have independent dynamics) but are assumed to be all equally connected to each other, (2) spatially The pioneer metapopulation model was first explicit approaches in which it is assumed that proposed by Levins (1969) (for detail see Box 1). local populations are arranged as cells on a regular This model has been modified and widely used in grid (lattice), with population sizes modelled as population studies because the habitats used by either discrete or continuous variables and where most species are becoming fragmented; therefore local populations are assumed to interact only the populations fall into metapopulation concepts with local populations in the nearby cells, and (3) (Box 2). spatially realistic approaches in which the models From the box 1, at equilibrium, dP = 0, the allow one to include in the model the specific dt geometry of particular patch networks, such as proportion of occupied sites can be investigated how many patches there are in the network, how by, large they are, and where exactly they are located. P* = 1 - e (1) For more details see Table 1. β Songklanakarin J. Sci. Technol. Metapopulation & Conservation Biology Vol. 25 No. 3 May-Jun. 2003 398 Yuttham, K., et al. Table 1. Three types of modelling approaches used in metapopulation studies. The table provides definitions, examples, advantages, and disadvantages of each type of modelling approach. Modelling Approaches Modelling Definition Examples Advantages Disadvantages Approaches Spatially Implicit Model in which all local Classical Levins model, It greatly facilitates It can be used to study only Models populations are equally and all patch models the mathematical and a subset of all interesting connected. and structured models conceptual analysis. questions, such as, what if migration rate is high enough to rescue local population? Spatially Explicit Model in which migration Cellular automata The mathematical The state of the metapopulation Models is distance-dependent, models, interacting rules that govern cannot be described simply by often restricted to the particle systems, and local behaviour are the fraction of cells occupied; nearest habitat patches; couple map lattice the same from cell to an entire vector of presences the patches are typically models cell and it is easy to and absences is needed. identical cells on a regular write a computer pro- grid, and only presence or gram to model the absence of the species in dynamics. a cell is considered. Spatially Model that assigns parti- Simulation models and