1: %\documentclass[floatfix,preprint,showpacs,showemail,preprintnumbers,amsfonts,pre]{revtex4}
2: \documentclass[showpacs,preprintnumbers,pre]{revtex4}
3: \usepackage{graphicx}% Include figure files
4: %\usepackage{dcolumn}% Align table columns on decimal point
5: \usepackage{bm}% bold math
6: %\usepackage{epsf}
7: %\usepackage[T1]{fontenc}
8: %\def\baselinestretch{2.0}
9: \pagestyle{plain}
10:
11:
12: \begin{document}
13:
14: \title{COMPLEX BEHAVIOR IN SIMPLE MODELS OF BIOLOGICAL COEVOLUTION
15: %\footnote{For the title, try not to use more than
16: %3 lines. Typeset the title in 10 pt roman, uppercase and
17: %boldface.}
18: }
19:
20: \author{PER ARNE RIKVOLD
21: %\footnote{
22: %Typeset names in 8 pt roman, uppercase. Use the footnote to indicate the
23: %present or permanent address of the author.}
24: }\email{rikvold@scs.fsu.edu}
25:
26: \affiliation{School of Computational Science, Center for Materials
27: Research and Technology,\\
28: Department of Physics, and National High Magnetic Field Laboratory\\
29: Florida State University, Tallahassee, Florida 32306-4120, USA\\
30: %\footnote{State completely without abbreviations, the
31: %affiliation and mailing address, including country. Typeset in 8 pt
32: %italic.}\\
33: }
34:
35:
36: \begin{abstract}
37: We explore the complex dynamical behavior of simple predator-prey
38: models of biological coevolution that account
39: for interspecific and intraspecific competition for resources, as well
40: as adaptive foraging behavior.
41: In long kinetic Monte Carlo simulations of these models we
42: find quite robust $1/f$-like noise in species diversity and
43: population sizes, as well as power-law distributions for the
44: lifetimes of individual species and the durations of quiet
45: periods of relative evolutionary stasis.
46: In one model, based on the Holling Type II functional response,
47: adaptive foraging produces a metastable low-diversity
48: phase and a stable high-diversity phase.
49: \end{abstract}
50:
51: \pacs{
52: 87.23.Kg, %Dynamics of evolution (In 80: Interdisciplinary physics.)
53: 05.40.-a, %Fluctuation phenomena, random processes, noise, and Brownian motion
54: 05.65.+b %Self-organized systems
55: }
56:
57: \maketitle
58:
59: \section{Introduction}
60: \label{sec:int}
61:
62: Biological evolution presents a rich array of phenomena that
63: involve nonlinear interactions between large numbers of units. As a
64: consequence, problems in evolutionary biology have recently
65: enjoyed increasing popularity among statistical and computational
66: physicists.\cite{DROS01} However, many of the models used by
67: physicists have unrealistic features that prevent the results from
68: attracting significant attention from biologists. In this
69: paper we therefore develop and explore individual-based models
70: of coevolution in predator-prey systems based on more
71: realistic population dynamics than some earlier
72: models.\cite{HALL02,CHRI02,COLL03,RIKV03,RIKV03A,ZIA04,RIKV05A,SEVI05,SEVI06,RIKV06}
73:
74:
75: \section{Models}
76: \label{sec:mod}
77:
78: Recently the author, together with R.~K.~P.\ Zia,
79: introduced a simplified form of the tangled-nature
80: model of biological macroevolution, which was developed by
81: Jensen and collaborators.\cite{HALL02,CHRI02,COLL03}
82: In these simplified
83: models,\cite{RIKV03,RIKV03A,ZIA04,RIKV05A,SEVI05,SEVI06,RIKV06}
84: the reproduction rates in an individual-based population
85: dynamics with nonoverlapping generations provide the mechanism for
86: selection between several interacting species. New species are
87: introduced into the community through point
88: mutations in a haploid, binary ``genome" of
89: length $L$, as in Eigen's model for molecular
90: evolution.\cite{EIGE71,EIGE88}
91: The potential species are identified by the index $I \in [0,2^L-1]$.
92: (Typically, only $\mathcal{N}(t) \ll 2^L$
93: of these species are present in the community at any one time $t$.)
94: At the end of each generation, each individual of species $I$ gives
95: birth to a fixed number
96: $F$ of offspring with probability $P_I$ before dying, or dies
97: without offspring with probability $(1-P_I)$. Each offspring may
98: mutate into a different species -- generally with different properties
99: -- with a small probability $\mu$. Mutation consists in flipping a
100: randomly chosen bit in the genome.
101:
102: \subsection{Simplified tangled-nature models}
103: \label{sec:tana}
104:
105: In these models, the reproduction probability for an individual of
106: species $I$ is given by the nonlinear function
107: \begin{equation}
108: P_I(t) = \frac{1}{1 + \exp[-\Delta_I(R,\{n_J(t)\})]} \;,
109: \label{eq:PI}
110: \end{equation}
111: where $R$ is an external resource that is renewed at the
112: same level each generation, and $\{n_J(t)\}$ is the set of
113: population sizes of all the species resident in the community in
114: generation $t$. The function $\Delta_I$ is given by
115: \begin{equation}
116: \Delta_I(R,\{n_J(t)\}) = - b_I + \eta_I R /N_{\rm tot}(t)
117: + \sum_J M_{IJ} n_J(t)/N_{\rm tot}(t) - N_{\rm tot}(t)/N_0
118: \;.
119: \label{eq:Delta}
120: \end{equation}
121: Here $b_I$ is an ``energy cost" of reproduction (always
122: positive), and $\eta_I$ (positive for primary producers or autotrophs,
123: and zero for consumers or heterotrophs) is the ability of
124: individuals of species $I$ to utilize the external resource $R$,
125: while $N_0$ is an environmental carrying capacity\cite{MURR89}
126: (a.k.a.\ Verhulst factor\cite{VERH1838}).
127: %which prevents the population size from diverging to infinity.
128: The total population size is $N_{\rm tot}(t) = \sum_J n_J(t)$.
129: The main feature of this reproduction probability is the random {\it
130: interaction matrix\/} $\bf M$,\cite{SOLE96} which is constructed at
131: the beginning of a simulation run, and thereafter kept constant
132: (quenched randomness). If $M_{IJ}$ is positive and $M_{JI}$
133: negative, $I$ is a predator and $J$ its prey, and vice versa. If
134: both matrix elements are positive, the relationship is a
135: mutualistic one, while both negative indicate an antagonistic
136: relationship.
137:
138: Two versions of this model were studied in earlier work. In the
139: first version, which we have called Model A,
140: there is no external resource or birth cost, and the off-diagonal
141: elements of $\bf M$ are stochastically independent and uniformly
142: distributed over $[-1,+1]$, while the diagonal elements are zero.
143: This model evolves toward mutualistic communities, in which all
144: species are connected by mutually
145: positive interactions.\cite{RIKV03,RIKV03A,ZIA04,SEVI05,SEVI06}
146:
147: Of greater biological interest is a predator-prey version of the
148: model, called Model B. In this case a small minority
149: of the potential species (typically 5\%) are primary producers, while the
150: rest are consumers. The off-diagonal part of the interaction matrix is
151: antisymmetric, with the additional restriction that a producer
152: cannot also prey on a consumer.\cite{RIKV05A,RIKV06}
153: In simulations we have taken $b_I$ and the nonzero $\eta_I$ as
154: independent and uniformly distributed on $(0,+1]$.
155: This model generates simple food webs with up to three
156: trophic levels.\cite{RIKV05A,RIKV06,RIKV06B}
157:
158: Both of these models provide interesting results, which include
159: intermittent dynamics with power spectral densities (PSDs) of diversities
160: and population sizes that exhibit $1/f$-like noise,
161: as well as power-law distributions for the lifetimes of individual
162: species and the duration of quiet periods of relative evolutionary stasis.
163: From a theoretical point of view they
164: also have the great advantage that the mean-field equation for the
165: steady-state
166: average population sizes in the absence of mutations reduces to a
167: set of linear (if $N_0 = \infty$) or at most quadratic equations and thus
168: can easily be solved exactly.\cite{RIKV03,ZIA04,RIKV06,RIKV06B}
169: The models thus provide useful benchmarks for more realistic,
170: but generally highly nonlinear models.
171:
172: In fact, the population dynamics
173: defined by Eqs.~(\ref{eq:PI}) and~(\ref{eq:Delta}) are not very
174: realistic. In particular, by summing over positive and negative
175: terms in $\Delta_I$, the models enable species with little food to
176: remain near a steady state if they are also not very popular as
177: prey, or have very low birth cost.
178: A more serious problem is the {\it ad-hoc\/} nature of the
179: normalization by the total population size $N_{\rm tot}(t)$ in the
180: resource and interaction terms in $\Delta_I$. While this is the source
181: of the models' analytic solvability, it implies an
182: indiscriminate, universal competition without regard to whether or
183: not two species directly utilize the same resources or share a
184: common predator. The purpose of the present paper is to develop models
185: with more realistic population dynamics and explore the
186: properties of their evolutionary dynamics.
187:
188: \subsection{Functional-response model}
189: \label{sec:holl}
190:
191: Here we develop a model with more realistic population
192: dynamics that include competition between different
193: predators that prey on the same species, as well as a saturation
194: effect expected to occur for a predator with abundant prey.
195: In doing so, we retain from the models discussed above the
196: important role of the interaction matrix $\bf M$, as well as the
197: restriction to dynamics with nonoverlapping generations.
198:
199: We first deal with the competition between predator species
200: by defining the number of
201: individuals of $J$ that are available as prey for $I$, corrected for
202: competition from other predator species, as
203: \begin{equation}
204: \hat{n}_{IJ} = \frac{n_I M_{IJ}}{\sum_L^{{\rm pred}(J)} n_L M_{LJ}} n_J \;,
205: \label{eq:neff}
206: \end{equation}
207: where $\sum_L^{{\rm pred}(J)}$ runs over all $L$
208: such that $M_{LJ} > 0$, i.e.,
209: over all predators of $J$. Thus,
210: $\sum_I^{{\rm pred}(J)} \hat{n}_{IJ} = n_J$, and if $I$ is the only
211: predator consuming $J$, then $\hat{n}_{IJ} = n_J$.
212:
213: Analogously, we define the competition-adjusted
214: external resources available to a producer species $I$ as
215: \begin{equation}
216: \hat{R}_I = \frac{n_I \eta_I}{\sum_L n_L \eta_L} R \;.
217: \label{eq:reff}
218: \end{equation}
219: As in the case of predators, $\sum_I \hat{R}_I = R$, and a
220: sole producer species has all of the external resources available to
221: it: $\hat{R}_I = R$. With these definitions, the total,
222: competition-adjusted resources available for the sustenance of
223: species $I$ are
224: \begin{equation}
225: \hat{S}_I = \eta_I \hat{R}_I + \sum_J^{{\rm prey}(I)} M_{IJ} \hat{n}_{IJ}
226: \;,
227: \label{eq:SI}
228: \end{equation}
229: where $\sum_J^{{\rm prey}(I)}$ runs over all $J$ such that $M_{IJ} >
230: 0$, i.e., over all prey of $I$.
231:
232: A central concept of the model is the {\it functional response\/} of
233: species $I$ with respect to $J$, $\Phi_{IJ}$.\cite{DROS01B,KREB01}
234: This is the rate at which an individual of species $I$ consumes
235: individuals of $J$. The simplest functional response corresponds to
236: the Lotka-Volterra model:\cite{MURR89} $\Phi_{IJ} = n_J$ if $M_{IJ} > 0$
237: and 0 otherwise. However, it is reasonable to expect that the
238: consumption rate should saturate in the presence of very abundant
239: prey.\cite{KREB01} For ecosystems consisting of a single pair of
240: predator and prey, or a simple chain reaching from a bottom-level
241: producer through intermediate species to a top predator, the most
242: common forms of functional response are due to Holling.\cite{KREB01}
243: For more complicated, interconnected food webs, a number of
244: functional forms have been proposed in the recent
245: literature,\cite{DROS01B,SKAL01,KUAN02,DROS04,MART06}
246: but there is as yet no agreement about a standard form.
247: Here we choose a ratio-dependent\cite{ABRA94,RESI95}
248: Holling Type II form,\cite{KREB01}
249: \begin{equation}
250: \Phi_{IJ} = \frac{M_{IJ} \hat{n}_{IJ}}{\lambda \hat{S}_I + n_I} \;,
251: \label{eq:PhiIJ}
252: \end{equation}
253: where $\lambda \in (0,1]$ is the metabolic efficiency of converting
254: prey biomass to predator offspring.
255: Analogously, the functional response of a producer species
256: $I$ toward the external resource $R$ is
257: \begin{equation}
258: \Phi_{IR} = \frac{\eta_I \hat{R}_{I}}{\lambda \hat{S}_I + n_I} \;.
259: \label{eq:PhiIR}
260: \end{equation}
261: In both cases, if $\lambda \hat{S}_I \ll n_I$, then the consumption
262: rate equals the resource ($M_{IJ} \hat{n}_{IJ}$ or
263: $\eta_I \hat{R}_{I}$) divided by the number of individuals of $I$,
264: thus expressing intraspecific competition for scarce
265: resources. In the opposite limit, $\lambda \hat{S}_I \gg n_I$,
266: the consumption rate is proportional to the ratio of the specific,
267: competition-adjusted resource to the competition-adjusted total
268: available sustenance, $\hat{S}_I$. The total consumption rate for
269: an individual of $I$ is therefore
270: \begin{equation}
271: C_I = \Phi_{IR} + \sum_J^{{\rm prey}(I)} \Phi_{IJ}
272: = \frac{\hat{S}_I}{\lambda
273: \hat{S}_I + n_I}
274: =
275: \left\{
276: \begin{array}{lll}
277: \hat{S}_I/n_I & \mbox{for} & \lambda \hat{S}_I \ll n_I \nonumber\\
278: 1/\lambda & \mbox{for} & \lambda \hat{S}_I \gg n_I
279: \end{array}
280: \right.
281: \;.
282: \label{eq:CI}
283: \end{equation}
284: The birth probability
285: is assumed to be proportional to the consumption rate,
286: \begin{equation}
287: B_I = \lambda C_I \in [0,+1] \;,
288: \label{eq:BI}
289: \end{equation}
290: while the probability that an individual of $I$
291: avoids death by predation until attempting to reproduce is
292: \begin{equation}
293: A_I = 1 - \sum_J^{{\rm pred}(I)} \Phi_{JI} \frac{n_J}{n_I} \;.
294: \label{eq:AI}
295: \end{equation}
296: The total reproduction probability for an individual of species $I$
297: in this model is thus $P_I(t) = A_I(t) B_I(t)$.
298:
299: \section{Numerical Results for the Functional-response Model}
300: \label{sec:Sim1}
301:
302: We simulated the functional-response model over
303: $2^{24} = 16\,777\,216$ generations (plus $2^{20}$
304: generations ``warm-up") for the
305: following parameters: genome length $L=21$
306: ($2^{21} = 2\,097\,152$ potential
307: species), external resource $R=16\,000$, fecundity $F=2$,
308: mutation rate $\mu = 10^{-3}$, proportion of
309: producers $c_{\rm prod} =0.05$, interaction matrix $\bf M$ with
310: connectance $C = 0.1$ and nonzero elements with a symmetric,
311: triangular distribution over $[-1,+1]$, and $\lambda = 1.0$.
312: We ran five independent runs, each starting from 100 individuals
313: of a single, randomly chosen producer species.
314:
315: \subsection{Time series}
316: \label{sec:timser}
317:
318: \begin{figure}[t]
319: \begin{center}
320: \vspace*{0.3truecm}
321: \includegraphics[angle=0,width=.47\textwidth]{timserNRC_B3fig.eps}
322: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/timserNRC_B3fig.eps
323: \hspace{0.5truecm}
324: \includegraphics[angle=0,width=.47\textwidth]{Creation-ExtinctionFig_NRCB3.eps}
325: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/Creation-ExtinctionFig_NRCB3..eps
326: \end{center}
327: \caption[]{
328: (Color online.)
329: {\bf (a)}
330: Time series of diversities (measured in species, lower curves)
331: and population sizes (measured in individuals, upper curves) for
332: one specific simulation run. The
333: strongly fluctuating curves in the background are sampled every
334: 8192 generations, while the smooth curves with data points in
335: contrasting colors that are overlaid
336: in the foreground are running averages over
337: 524\,288 generations. Black with light gray (yellow online)
338: overlay: all species.
339: Light gray with dark gray overlay (green and magenta online): producers.
340: Dark gray with light gray overlay (red and cyan online): consumers.
341: {\bf (b)}
342: Time series of the
343: accumulation of new and extinct species in the same simulation run
344: depicted in (a).
345: The solid, black curve shows the total number of different
346: species that have at least once
347: attained a population size $n_I > 1000$ by time $t$. The dashed curves
348: count the total number of species that have gone extinct after
349: attaining a maximum population greater than 1000. The black dashed
350: curve refers to all species, the light gray one (green online) to
351: producers, and the dark gray one (red online) to consumers. The
352: ratio of approximately 1.89 between the dashed and full black curves
353: indicate that major species recur on average about twice
354: during the evolution. This is an artifact of the finite genome
355: length. The inset shows the detailed, intermittent structure of
356: the solid, black curve over $400\,000$ generations. The interval is
357: indicated by a horizontal bar in the main panel.
358: }
359: \label{fig:timser}
360: \end{figure}
361: Time series of diversities (effective numbers of species)
362: and population sizes
363: for one realization are shown in Fig.~\ref{fig:timser}. To
364: filter out noise from low-population,
365: unsuccessful mutations, we define the diversity as the exponential
366: Shannon-Wiener index.\cite{KREB89} This is the exponential function
367: of the information-theoretical entropy of the population distributions,
368: $D(t) = \exp \left[S \left( \{ n_I(t) \} \right) \right]$, where
369: \begin{equation}
370: S\left( \{ n_I(t) \} \right)
371: =
372: - \sum_{\{I | \rho_I(t) > 0 \}} \rho_I(t) \ln \rho_I(t)
373: \label{eq:S}
374: \end{equation}
375: with
376: $\rho_I(t) = n_I(t) / N_{\rm tot}(t)$ for the case of all species,
377: and analogously for the producers and consumers separately.
378:
379: The time series for both diversities and population sizes show
380: intermittent behavior with quiet periods of varying lengths,
381: separated by periods of high evolutionary
382: activity. In this respect, the results
383: are similar to those seen for Models A and B in earlier
384: work.\cite{RIKV03,RIKV03A,RIKV05A,SEVI06,RIKV06} However,
385: diverse communities in this model seem to be less stable than those
386: produced by the linear models. In particular, this model has a
387: tendency to flip randomly between an active
388: phase with a diversity near
389: ten, and a ``garden of Eden'' phase of one or a few producers with
390: a very low population of unstable consumers, such as the one seen
391: around 10 million generations in Fig.~\ref{fig:timser}.
392:
393:
394: \subsection{Power-spectral densities}
395: \label{sec:psd}
396:
397: \begin{figure}[t]
398: \begin{center}
399: \vspace*{0.1truecm}
400: \includegraphics[angle=0,width=.47\textwidth]{PSDdiv-popFigNRCB.eps}
401: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/PSDdiv-popFigNRCB.eps
402: \hspace{0.5truecm}
403: \includegraphics[angle=0,width=.47\textwidth]{PSDextFigNRCB.eps}
404: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/PSDextFigNRCB.eps
405: \end{center}
406: \caption[]{
407: (Color online.)
408: {\bf (a)}
409: PSDs for the diversities and population sizes, each recorded
410: separately for all species and for producers and consumers.
411: The time series were sampled every 16 generations.
412: {\bf (b)}
413: PSDs for the extinction activity.
414: In both parts of the figure, the results are averaged over five
415: independent simulation runs.
416: See discussion in the text.
417: }
418: \label{fig:psd}
419: \end{figure}
420: A common method to obtain information about the intensity of
421: fluctuations in a time series at different time scales is the
422: power-spectral density (squared Fourier transform), or PSD. PSDs are
423: presented in Fig.~\ref{fig:psd} for the diversity fluctuations and the
424: fluctuations in the population sizes (Fig.~\ref{fig:psd}(a))
425: and the intensity of extinction events (Fig.~\ref{fig:psd}(b)).
426: The former two are shown for the total population, as well as
427: separately for the
428: producers and consumers. All three are similar.
429: Extinction events are recorded as
430: the number of species that have attained a population size greater
431: than one, which go extinct in generation $t$ (marked as ``species"
432: in the figure), while
433: extinction sizes are calculated by adding the maximum populations
434: attained by all species that go extinct in generation $t$ (marked
435: as ``population" in the figure).
436: The PSDs for all the quantities shown exhibit approximate $1/f$
437: behavior. For the diversities and population sizes, this power law
438: extends over more than five decades in time. The extinction
439: measures, on the other hand, have a large background of white noise
440: for frequencies above $10^{-3}$ generations$^{-1}$,
441: probably due to the high rate of
442: extinction of unsuccessful mutants. For lower frequencies, however,
443: the behavior is consistent with $1/f$ noise within the limited
444: accuracy of our results.
445:
446: \subsection{Species lifetimes and durations of quiet periods}
447: \label{sec:times}
448:
449: \begin{figure}[t]
450: \begin{center}
451: \vspace*{1.1truecm}
452: \includegraphics[angle=0,width=.47\textwidth]{SpecLifeFigNRCB.eps}
453: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/SpecLifeFigNRCB.eps
454: \hspace{0.5truecm}
455: \includegraphics[angle=0,width=.47\textwidth]{PerDurdSdtFig.eps}
456: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/PerDurdSdtFig.eps
457: \end{center}
458: \caption[]{
459: (Color online.)
460: {\bf (a)}
461: Histograms of species lifetimes, shown for all species, as well as
462: separately for producers and consumers.
463: {\bf (b)}
464: Histograms of the durations of evolutionarily quiet periods,
465: defined as the times that $|{\rm d}S/{\rm d}t|$ (averaged over 16 generations)
466: falls continuously below some cutoff. The inset is a histogram of
467: ${\rm d}S/{\rm d}t$, showing a Gaussian center with approximately exponential
468: wings. The parabola in the foreground is a Gaussian fit to this
469: central peak. The cutoff values for the main figure,
470: between 0.008 and 0.024, were chosen on the basis of this distribution.
471: The data in both parts of the figure
472: are averaged over five independent simulation runs.
473: }
474: \label{fig:time}
475: \end{figure}
476: The evolutionary dynamics can also be characterized by histograms
477: of characteristic time intervals, such as the time from creation till
478: extinction of a species (species lifetimes) or the time intervals
479: during which
480: some indicator of evolutionary activity remains continuously below
481: a chosen cutoff (duration of evolutionarily quiet periods).
482: Histograms of species lifetimes are shown in Fig.~\ref{fig:time}(a).
483: As our indicator of evolutionary activity we use the magnitude of the
484: logarithmic derivative of the diversity, $|{\rm d}S/{\rm d}t|$,
485: and histograms for the resulting durations of
486: quiet periods, calculated with different cutoffs, are shown in
487: Fig.~\ref{fig:time}(b). Both quantities display approximate
488: power-law behavior with an exponent near $-2$, consistent with the
489: $1/f$ behavior observed in the PSDs.\cite{RIKV03,PROC83}
490: It is interesting to note that the distributions for these two
491: quantities for this model have approximately the same exponent.
492: This is consistent with the previously studied, mutualistic
493: Model A,\cite{RIKV03,RIKV05A}
494: but not with the predator-prey Model B.\cite{RIKV05A,RIKV06,RIKV06B}
495: We believe the linking of the power laws for the species lifetimes
496: and the duration of quiet periods indicate that the communities
497: formed by the model are relatively fragile, so that all member
498: species tend to go extinct together in a ``mass extinction." In
499: contrast, Model B produces simple food webs that are much more
500: resilient against the loss of a few species, and as a result the
501: distribution of quiet-period durations decays with an exponent near
502: $-1$.\cite{RIKV05A,RIKV06,RIKV06B}
503:
504:
505: \section{Adaptive Foraging}
506: \label{sec:adap}
507:
508: \begin{figure}[t]
509: \begin{center}
510: \vspace*{0.2truecm}
511: \includegraphics[angle=0,width=.47\textwidth]{timsADFig.eps}
512: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/timsADFig.eps
513: \hspace{0.5truecm}
514: \includegraphics[angle=0,width=.47\textwidth]{Creation-ExtinctionFig_NRC_AD1.eps}
515: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/Creation-ExtinctionFig_NRC_AD1.eps
516: \end{center}
517: \caption[]{
518: (Color online.)
519: {\bf (a)}
520: Time series of
521: population sizes (upper curves) and diversities (lower curves) for the model
522: with adaptive foraging. The interpretation of the colors and lines are
523: the same as in Fig.~\protect\ref{fig:timser}(a).
524: {\bf (b)}
525: Time series of
526: the number of new species that have reached a population greater than
527: 1000 (lower curve) and greater than 100 (upper curve). The inset shows
528: the intermittent structure of the upper curve on a very fine scale of
529: 2000 generations. See discussion in the text.
530: }
531: \label{fig:adap}
532: \end{figure}
533: The model studied above is one in which species forage
534: indiscriminately over all available resources, with the output only
535: limited by competition. Also, there is an implication that an
536: individual's total foraging effort increases proportionally with
537: the number of species to which it is connected by a positive
538: $M_{IJ}$. A more realistic picture would be that an individual's
539: total foraging effort is constant and can either be divided
540: equally, or concentrated on richer resources. This is known as
541: adaptive foraging. While one can
542: go to great length devising optimal foraging
543: strategies,\cite{DROS01B,DROS04} we here
544: only use a simple scheme, in which individuals of $I$ show a
545: preference for prey species $J$, based on the interactions and
546: population sizes (uncorrected for interspecific competition) and given by
547: \begin{equation}
548: g_{IJ} = \frac{M_{IJ}n_J}{\eta_I R + \sum_K^{{\rm prey}(I)} M_{IK} n_K}
549: \;,
550: \label{eq:gij}
551: \end{equation}
552: and analogously for $R$ by
553: \begin{equation}
554: g_{IR} = \frac{\eta_{I} R}{\eta_I R + \sum_K^{{\rm prey}(I)} M_{IK} n_K}
555: \;.
556: \label{eq:gir}
557: \end{equation}
558: The total foraging effort is thus
559: $g_{IR} + \sum_J^{{\rm prey}(I)} g_{IJ} = 1$.
560: The preference factors are used to modify the reproduction
561: probabilities by replacing all occurrences of $M_{IJ}$
562: by $M_{IJ} g_{IJ}$ and of $\eta_I$ by $\eta_I g_{IR}$ in
563: Eqs.~(\ref{eq:neff}--\ref{eq:PhiIR}).
564:
565: The results of implementing the adaptive foraging are quite striking.
566: The system appears now to have a metastable low-diversity phase
567: similar to the active phase of the non-adaptive model, from
568: which it switches at a random time to an apparently stable
569: high-diversity phase with much smaller fluctuations.
570: As seen in Fig.~\ref{fig:adap}(a), the switchover
571: is quite abrupt, and Fig.~\ref{fig:adap}(b) shows that it is accompanied by
572: a sudden reduction in the rate of creation of new species.
573: As seen in Fig.~\ref{fig:adapPSD}, the PSDs for both the diversities and
574: population sizes in both phases show approximate $1/f$ noise for
575: frequencies above $10^{-5}$ generations$^{-1}$. For lower
576: frequencies, the metastable phase shows no discernible frequency
577: dependence, while for the stable phase, the frequency dependence continues
578: at least another decade. It thus appears that long-time correlations are
579: not seen beyond $10^5$ generations for the metastable phase, and
580: probably not beyond about $10^6$ generations for the stable one.
581: These observations are
582: consistent with species-lifetime distributions for both phases
583: (not shown), which are quite similar to those for the
584: non-adaptive model, but typically with cutoffs in the range of $10^5$ to
585: $10^6$ generations, much shorter than the total simulation times.
586:
587: \begin{figure}[t]
588: \begin{center}
589: \vspace*{0.2truecm}
590: \includegraphics[angle=0,width=.47\textwidth]{PSDdiv-popFigAD_METAST.eps}
591: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/ADAP/xxx
592: \hspace{0.5truecm}
593: \includegraphics[angle=0,width=.47\textwidth]{PSDdiv-popFigAD_STABLE.eps}
594: %/home/r3/rikvold/scratch/Modified/Predprey/LargeL/LARGE/HOLLING/ADAP/xxx
595: \end{center}
596: \caption[]{
597: (Color online.)
598: PSDs for the diversities and population sizes in the metastable phase
599: (averaged over three independent runs)
600: {\bf (a)}, and the stable phase (averaged over five independent runs)
601: {\bf (b)} for the model with adaptive foraging. Both show
602: approximate $1/f$ noise for frequencies above about $10^{-5}$,
603: but the PSDs appear to approach constant levels for the lowest
604: frequencies.
605: }
606: \label{fig:adapPSD}
607: \end{figure}
608: In fact, the system can also escape from the
609: low-diversity phase to total extinction, which is an absorbing state,
610: and in some of our simulation runs we avoided this by limiting
611: $|M_{IJ}|$ to less than 0.9. This restriction does not seem to have any
612: effect on the dynamics in the high-diversity phase.
613: These results are preliminary, and it is possible that the
614: high-diversity phase corresponds to a mutational meltdown. More research
615: is clearly needed regarding the effects of adaptive foraging in this
616: model.
617:
618:
619: \section{Conclusions}
620: \label{sec:conc}
621:
622: In this paper we have shown that very complex and diverse dynamical
623: behavior results, even from highly over-simplified models of biological
624: macroevolution. In particular, PSDs that show $1/f$-like noise and
625: power-law lifetime distributions for species as well as evolutionarily quiet
626: states are generally seen. This is the case, both in the analytically
627: tractable, but somewhat unrealistic tangled-nature type models, and in
628: the nonlinear predator-prey models based on the more realistic Holling Type II
629: functional response. Particularly intriguing is the appearance of a new,
630: stable high-diversity phase in the latter type of model when adaptive
631: foraging behavior is included. Among the many questions about this new
632: phase that remain to be addressed is the structure of the resulting
633: community food webs.
634:
635:
636: \section*{Acknowledgments}
637:
638: Supported in part by U.S.\ National Science Foundation Grant Nos.\
639: DMR-0240078 and DMR-0444051 and by Florida State University through the
640: School of Computational Science, the Center for Materials Research
641: and Technology, and the National High Magnetic Field Laboratory.
642:
643:
644: %\begin{thebibliography}{000} %for 3 digits
645: %\begin{thebibliography}{00} %for 2 digits
646: %\begin{thebibliography}{0} %for 1 digit
647:
648: %\end{thebibliography}
649:
650: %\bibliography{evol.bib}
651: %\bibliographystyle{prsty}
652:
653:
654: \begin{thebibliography}{10}
655:
656: \bibitem{DROS01}
657: B. Drossel, Adv.\ Phys. {\bf 50}, 209 (2001).
658:
659: \bibitem{HALL02}
660: M. Hall, K. Christensen, S.~A. di~Collobiano, and H.~J. Jensen, Phys.\ Rev.\ E
661: {\bf 66}, 011904 (2002).
662:
663: \bibitem{CHRI02}
664: K. Christensen, S.~A. di~Collobiano, M. Hall, and H.~J. Jensen, J.\ theor.\
665: Biol. {\bf 216}, 73 (2002).
666:
667: \bibitem{COLL03}
668: S.~A. di~Collobiano, K. Christensen, and H.~J. Jensen, J.\ Phys.\ A {\bf 36},
669: 883 (2003).
670:
671: \bibitem{RIKV03}
672: P.~A. Rikvold and R.~K.~P. Zia, Phys.\ Rev.\ E {\bf 68}, 031913 (2003).
673:
674: \bibitem{RIKV03A}
675: P.~A. Rikvold and R.~K.~P. Zia, in {\em Computer Simulation Studies in
676: Condensed Matter Physics XVI}, edited by D.~P. Landau, S.~P. Lewis, and H.-B.
677: Sch{\"u}ttler (Springer-Verlag, Berlin, 2004), pp.\ 34--37.
678: %, e-print arXiv:nlin.AO/0303010.
679:
680: \bibitem{ZIA04}
681: R.~K.~P. Zia and P.~A. Rikvold, J.\ Phys.\ A {\bf 37}, 5135 (2004).
682:
683: \bibitem{RIKV05A}
684: P.~A. Rikvold, in {\em Noise in Complex Systems and Stochastic Dynamics III},
685: edited by L.~B. Kish, K. Lindenberg, and Z. Gingl (SPIE, The International
686: Society for Optical Engineering, Bellingham, WA, 2005), pp.\ 148--155,
687: e-print arXiv:q-bio.PE/0502046.
688:
689: \bibitem{SEVI05}
690: V. Sevim and P.~A. Rikvold, in {\em Computer Simulation Studies in Condensed
691: Matter Physics XVII}, edited by D.~P. Landau, S.~P. Lewis, and H.-B.
692: Sch{\"u}ttler (Springer-Verlag, Berlin, 2005), pp.\ 90--94.
693: %, e-print arXiv:q-bio.PE/0403042.
694:
695: \bibitem{SEVI06}
696: V. Sevim and P.~A. Rikvold, J.\ Phys.\ A {\bf 38}, 9475 (2005).
697:
698: \bibitem{RIKV06}
699: P.~A. Rikvold, arXiv:q-bio.PE/0508025.
700:
701: \bibitem{EIGE71}
702: M. Eigen, Naturwissenschaften {\bf 58}, 465 (1971).
703:
704: \bibitem{EIGE88}
705: M. Eigen, J. McCaskill, and P. Schuster, J.\ Phys.\ Chem. {\bf 92}, 6881
706: (1988).
707:
708: \bibitem{MURR89}
709: J.~D. Murray, {\em Mathematical Biology} (Springer-Verlag, Berlin, 1989).
710:
711: \bibitem{VERH1838}
712: P.~F. Verhulst, Corres.\ Math.\ et Physique {\bf 10}, 113 (1838).
713:
714: \bibitem{SOLE96}
715: R.~V. Sol{\'e} and J. Bascompte, Proc.\ R.\ Soc.\ Lond.\ B {\bf 263}, 161
716: (1996).
717:
718: \bibitem{RIKV06B}
719: P.~A. Rikvold, in preparation.
720:
721: \bibitem{DROS01B}
722: B. Drossel, P.~G. Higgs, and A.~J. McKane, J.\ theor.\ Biol. {\bf 208}, 91
723: (2001).
724:
725: \bibitem{KREB01}
726: C.~J. Krebs, {\em Ecology. The Experimental Analysis of Distribution and
727: Abundance. Fifth Edition.} (Benjamin Cummings, San Francisco, 2001).
728:
729: \bibitem{SKAL01}
730: G.~T. Skalski and J.~F. Gilliam, Ecology {\bf 82}, 3083 (2001).
731:
732: \bibitem{KUAN02}
733: Y. Kuang, J.\ Biomath. {\bf 17}, 129 (2002).
734:
735: \bibitem{DROS04}
736: B. Drossel, A. McKane, and C. Quince, J.\ theor.\ Biol. {\bf 229}, 539
737: (2004).
738:
739: \bibitem{MART06}
740: N.~D. Martinez, R.~J. Williams, and J.~A. Dunne, in {\em Ecological Networks:
741: Linking structure to dynamics in food webs}, edited by M. Pasqual and J.~A.
742: Dunne (Oxford University Press, Oxford, 2006), pp.\ 163--185.
743:
744: \bibitem{ABRA94}
745: P.~A. Abrams, Ecology {\bf 75}, 1842 (1994).
746:
747: \bibitem{RESI95}
748: H. Resit, R. Arditi, and L.~R. Ginzburg, Ecology {\bf 76}, 995 (1995).
749:
750: \bibitem{KREB89}
751: C.~J. Krebs, {\em Ecological Methodology} (Harper \& Row, New York, 1989),
752: chap.~10.
753:
754: \bibitem{PROC83}
755: I. Procaccia and H. Schuster, Phys.\ Rev.\ A {\bf 28}, 1210 (1983).
756:
757: \end{thebibliography}
758:
759: \end{document}
760: