Abstract
In this paper, we study the stationary and oscillatory Turing instabilities of a homogeneous equilibrium in prey–predator reaction–diffusion systems with dormant phase of predators. We propose a simple criterion which is useful in classifying these Turing instabilities. Moreover, numerical simulations reveal transient spatio-temporal complex patterns which are a mixture of spatially periodic steady states and traveling/standing waves. In this mixture, the steady part is the stable Turing pattern bifurcated primarily from the homogeneous equilibrium, while wave parts are unstable oscillatory solutions bifurcated secondarily from the same homogeneous equilibrium. Although our criterion does not exclude the occurrence of oscillatory Turing instability, we have not yet found stable traveling/standing waves due to oscillatory Turing instability in our simulations. These results suggest that dormancy of predators is not a generator but an enhancer of spatio-temporal Turing patterns in prey–predator reaction–diffusion systems.
Similar content being viewed by others
References
Anma A, Sakamoto K, Yoneda T (2012) Unstable subsystem cause Turing instability. Kodai Math J 35:215–247
Banerjee M (2010) Self-replication of spatial patterns in a ratio-dependent predator–prey model. Math Comp Model 51:44–52
Baurmann M, Gross T, Feudel U (2007) Instabilities in spatially extended predator-prey systems: spatio-temporal patterns in the neighborhood of Turing–Hopf bifurcations. J Theor Biol 245:220–229
Bazykin AD (1998) Nonlinear dynamics of interacting populations. World Scientific, Singapore
Bilinsky L, Hadeler KP (2009) Quiescence stabilizes predator–prey relations. J Biol Dyn 3:196–208
Cangelosi RA, Kealy-Dichone BJ, Wollkind DJ, Chaiya I (2014) Nonlinear stability analyses of Turing patterns for a mussel-algae model. J Math Biol (published online)
Cohen D, Levin SA (1987) The interaction between dispersal and dormancy strategies in varying and heterogeneous environments. In: Teramoto E, Yamaguchi M (eds) Mathematical topics in population biology, morphogenesis and neurosciences, Proceedings, Kyoto 1985, Lecture Notes in Biomath 71. Springer, pp 110–122
Cross GW (1978) Three types of matrix instability. Linear Alg Appl 20:253–263
Doedel EJ, Champneys AR, Dercole F, Fairgrieve TF, Kuznetsov YA, Oldeman BE, Paffenroth RC, Sandstede B, Wang X, Zhang C (2008) AUTO-07p: continuation and bifurcation software for ordinary differential equations
Garvie MR, Maini PK, Trenchea C (2010) An efficient and robust numerical algorithm for estimating parameters in Turing systems. J Comp Phys 229:7058–7071
Gyllström M, Hansson L-A (2004) Dormancy in freshwater zooplankton: induction, termination and the importance of benthic–pelagic coupling. Aquat Sci 66:274–295
Hairston NG Jr, Van Brunt RA, Kearns CM (1995) Age and survivorship of diapausing eggs in a sediment egg bank. Ecology 76:1706–1711
Kealy BJ, Wollkind DJ (2012) A nonlinear stability analysis of vegetative Turing pattern formation for an interaction–diffusion plant–surface water model system in an arid flat environment. Bull Math Biol 74:803–833
Kuwamura M (2005) On the Turing patterns in one-dimensional gradient/skew-gradient dissipative systems. SIAM J Appl Math 65:618–643
Kuwamura M, Chiba H (2009) Mixed-mode oscillations and chaos in a prey-predator system with dormancy of predators. Chaos 19:043121
Kuwamura M, Nakazawa T (2011) Dormancy of predators dependent on the rate of variation in prey density. SIAM J Appl Math 71:169–179
Kuwamura M, Nakazawa T, Ogawa T (2009) A minimum model of prey–predator system with dormancy of predators and the paradox of enrichment. J Math Biol 58:459–479
McGehee EA, Peacock-López E (2005) Turing patterns in a modified Lotka–Volterra model. Phys Lett A 342:90–98
Medvinsky AB, Petrovskii SV, Tikhonova IA, Malchow H, Li B-L (2002) Spatiotemporal complexity of plankton and fish dynamics. SIAM Rev 44:311–370
Murray JD (1989) Mathematical biology, 2nd edn. Springer, New York
Nakazawa T, Kuwamura M, Yamamura N (2011) Implications of resting eggs of zooplankton for the paradox of enrichment. Popul Ecol 53:341–350
Okubo A, Levin SA (2001) Diffusion and ecological problems, 2nd edn. Springer, New York
Rashevsky N (1938) Mathematical biophysics. Univ. of Chicago Press, Chicago
Satnoianu RA, Menzinger M, Maini PK (2000) Turing instabilities in general systems. J Math Biol 41:493–512
Sun G-Q, ** Z, Li L, Li B-L (2010) Self-organized wave pattern in a predator–prey model. Nonlinear Dyn 60:265–275
Turing AM (1952) The chemical basis of morphogenesis. Philos R Soc B 237:37–72
Upadhyaya RK, Volpert V, Thakura NK (2012) Propagation of Turing patterns in a plankton model. J Biol Dyn 6:524–538
Yang L, Dolnik M, Zhabotinsky AM, Epstein IR (2002) Pattern formation arising from interactions between Turing and wave instabilities. J Chem Phys 117:7259–7265
Wang L, Li MY (2001) Diffusion-driven instability in reaction–diffusion systems. J Math Anal Appl 254:138–153
White KAJ, Gilligan CA (1998) Spatial heterogenity in three species, plant–parasite–hyperparasite, systems. Philos Trans R Soc Lond B 353:543–557
Acknowledgments
The authors express his sincere gratitude to Professor Kunimochi Sakamoto for his useful advice and comments. Moreover, the author would like to appreciate the referees for their useful suggestions and comments, which have improved the original manuscript.
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Appendices
Appendix A: Proof of Theorem 3.1
Since \(A\) is stable, we have
and
Let
and
Direct calculation shows that
and
by virtue of (6.1) and (6.2). Moreover, we have \( \det B_{13} = - \alpha (\bar{u}) ( s'(\bar{u}) - f'(\bar{u})\bar{v} ), \det B_{23} = \alpha (\bar{u}) m'(\bar{v}) \bar{v} \) and
where
is independent of \(\alpha (\bar{u})\). Therefore, we have
by virtue of (6.1) and \( \alpha '(\bar{u}) \ge 0 \). Hence, \(p_1 p_2 - p_3 > 0\) holds for sufficiently small \(\alpha (\bar{u})\) if \(Y > 0\). Noting \(\mu (\bar{u}) + \nu (\bar{u}) = 1\), we have
When \( s'(\bar{u}) \le 0\), we have \(Y >0\). On the other hand, when \( s'(\bar{u}) > 0\), we have \( Y > 0\) under the condition (3.5) because \(\mu (\bar{u}) + \nu (\bar{u}) = 1\). Thus, \(p_1 p_2 - p_3 > 0\) holds for sufficiently small \(\alpha (\bar{u})\) under the condition (3.5). Therefore, it follows from the Routh–Hurwitz condition that \(B\) is stable if \(A\) is stable and (3.5) holds.
Appendix B: Proof of Theorem 3.3
By (Cross 1978, Theorem 4), Turing instability does not occur if and only if \(B\) is strongly stable, i.e., all the signed principal minors of \(B\) are nonnegative. Direct calculations show
where \(Y\) is given by (6.6). Noting (6.2) and that \(Y \ge 0\) if \( s'(\bar{u}) - f'(\bar{u})\bar{v} \le 0\), we see that \(B\) is strongly stable if and only if \( s'(\bar{u}) - f'(\bar{u})\bar{v} \le 0\).
Appendix C: Proof of Theorem 3.4
Since \(B\) is stable, it follows from the Routh–Hurwitz condition that
and
where \(p_1, p_2\) and \(p_3\) are given by (6.3). Notice that (8.1)–(8.3) imply
which plays a key role in the subsequent analysis.
We analyze the eigenvalue of \( -\sigma D_B + B \), where \(\sigma > 0\) and
The characteristic equation of \( -\sigma D_B + B \) is given by
where
Since \(p_1(\sigma ) > 0 \) for all \(\sigma >0\) by (8.1), it is impossible that both \(p_3(\sigma ) = 0\) and \(p_1(\sigma )p_2(\sigma ) - p_3(\sigma ) = 0\) hold due to the assumption of the simplicity of the eigenvalue \(\lambda = \lambda (\sigma , D_B; B)\) (see, Assumption 2.2). Therefore, it follows from the Routh–Hurwitz condition that a pair of complex eigenvalues crosses the imaginary axis and enters into the right half plane if the sign of \(p_1(\sigma )p_2(\sigma ) - p_3(\sigma )\) changes from positive to negative. Let
Lemma 8.1
When \( m'(\bar{v}) > 0\) and \( s'(\bar{u}) - f'(\bar{u})\bar{v} < \alpha (\bar{u}) \), there exists \(\delta _2 >0\) independent of \(\sigma \) such that \(F(\sigma ) > \delta _2\) holds for all \(\sigma >0 \) and for all \(d_u > 0\).
Proof
Let \(I(d_u)\) and \(J(d_u)\) denote the coefficients of \(\sigma \) and \(\sigma ^2\) in \(F(\sigma )\), respectively, which are given by
and
We will prove that \(I(d_u) >0 \) and \(J(d_u) >0 \) hold for all \(d_u >0 \). In order to show \(I(d_u) >0 \), we have to prove \(\det B_{12} + \det B_{23} + (\mathrm{tr} B)(\mathrm{tr} B_{13}) > 0\) and \(\det B_{12} + \det B_{13} + (\mathrm{tr} B)(\mathrm{tr} B_{23}) > 0\). Assume that \(\det B_{12} + \det B_{23} + (\mathrm{tr} B)(\mathrm{tr} B_{13}) \le 0\) holds. Then, we have
which is equivalent to
where \( X:= - \nu (\bar{u}) k f(\bar{u}) - m'(\bar{v})\bar{v},\,Y:=s'(\bar{u}) - f'(\bar{u})\bar{v},\,Z:= f(\bar{u}) \{ (\mu '(\bar{u})f(\bar{u}) + \mu (\bar{u})f'(\bar{u}) )k \bar{v} + \alpha '(\bar{u})\bar{w} \},\,\bar{\alpha } = \alpha (\bar{u}),\,\bar{\nu } = \nu (\bar{u})\) and \(\bar{f} = f(\bar{u})\). On the other hand, by (6.3) and (8.4), we have
Therefore, it follows from (8.8) and (8.9) that
which leads to a contradiction because \( Y < \bar{\alpha } \) and
by virtue of \(0 < \mu (\bar{u}) < 1,\,f(\bar{u}) >0\) and (6.1). Therefore, \(\det B_{12} + \det B_{23} + (\mathrm{tr} B)(\mathrm{tr} B_{13}) > 0\). Similarly, we can prove \(\det B_{12} + \det B_{13} + (\mathrm{tr} B)(\mathrm{tr} B_{23}) > 0\). Hence, we see that \(I(d_u) > 0\) holds.
Furthermore, we see that \(J(d_u) > 0\) holds because \(\mathrm{tr} B_{23} = X - \bar{\alpha } < 0,\,\mathrm{tr} B_{13} = Y - \bar{\alpha } < 0\), and \(\mathrm{tr} B = -p_1 < 0\) by (8.1).
Thus, noting \(\delta _2 := p_1p_2 -p_3 > 0\) by (8.3), we see that \(F(\sigma ) > \delta _2 \) holds for all \(\sigma >0 \) and for all \(d_u >0 \). \(\square \)
Lemma 8.2
When \( m'(\bar{v}) > 0\) and \(s'(\bar{u}) - f'(\bar{u})\bar{v} > \alpha (\bar{u}) \), there exists \(\delta _3 >0\) independent of \(\sigma \) such that \(F(\sigma ) > \delta _3\) holds for all \(\sigma >0 \) and for \(d_u \ge d^*\), where \(d^*\) is given by (3.7). Moreover, there exists \(\delta _4 >0\) independent of \(\sigma \) such that \(F(\sigma ) < 0\) holds for \( \sigma > \delta _4\) and for \(d_u =0\).
Proof
In order to show the first assertion of this lemma, it suffices to prove that \(I(d_u) >0 \) and \(J(d_u) >0 \) hold for \(d_u \ge d^*\), where \(I(d_u)\) and \(J(d_u)\) are given by (8.6) and (8.7), respectively. First, we will show that \(\det B_{12} + \det B_{13} + (\mathrm{tr} B)(\mathrm{tr} B_{23}) > 0\) holds. Assume that \(\det B_{12} + \det B_{13} + (\mathrm{tr} B)(\mathrm{tr} B_{23}) \le 0\), i.e.,
where \(X, Y, Z\) and \(\bar{\alpha }\) are the same notations in the proof of Lemma 8.1. By using the same line of arguments as applied to (8.8), it follows from (8.9) and (8.11) that
because \(X < kf(\bar{u}) - m(\bar{v}) < 0\) and \( X + Y < 0\) by (8.10), where \(\bar{\nu }\) and \(\bar{f}\) are the same notations in the proof of Lemma 8.1. Hence, we have \( \bar{\alpha } < X + Y - \bar{\nu } k \bar{f} < 0 \), which contradicts \(\bar{\alpha } > 0\). Therefore, \(\det B_{12} + \det B_{13} + (\mathrm{tr} B)(\mathrm{tr} B_{23}) > 0\).
Since \(I(d_u) >0\) holds for all \(d_u > 0\) if \(\det B_{12} + \det B_{23} + (\mathrm{tr} B)(\mathrm{tr} B_{13}) \ge 0\), we consider the case that \(\det B_{12} + \det B_{23} + (\mathrm{tr} B)(\mathrm{tr} B_{13}) <0\). In this case, \(I(d_u) = 0\) has the positive root \(d_1^*\) given by
We will prove that \( d_1^* \le d^*\) holds. Assume that \( d_1^* > d^*\) holds. Then, noting (3.7) and \(\mu (\bar{u}) + \nu (\bar{u}) = 1\), we have
which implies that
where \( W := \bar{\alpha }^2 - 2\bar{\alpha } (X+Y) + XY + Z\). On the other hand, by \(\det B_{12} + \det B_{23} + (\mathrm{tr} B)(\mathrm{tr} B_{13}) < 0\), we have \(W + Y(X+Y) - \bar{\alpha } \bar{\nu } k \bar{f} < 0\). Therefore, we have \(W(X-Y) - \bar{\alpha } \bar{\nu } k \bar{f} X > 0\). Since \(X- Y < X < 0\) by \(Y > \bar{\alpha } > 0\), we have
which implies
by virtue of (8.9). Hence, we have \( \bar{\alpha } < X + Y < 0\) by (8.10), which contradicts \(\bar{\alpha } >0 \). Thus, we see that \( d_1^* \le d^*\) holds, and hence \(I(d_u) > 0\) holds for \( d_u \ge d^*\).
Next, we will show that \(J(d_u) > 0\) holds for \( d_u \ge d^*\). Since \(\mathrm{tr} B_{23} = X - \bar{\alpha } < 0\) and \(\mathrm{tr} B_{13} = Y - \bar{\alpha } > 0\), we see that \(J(d_u) > 0\) holds for \( d_u > d_2^*\), where \(d_2^*\) is the positive root of \(J(d_u) = 0\). A direct calculation shows
Noting \(X - \bar{\alpha } < 0,\, Y >0\) and \( X + Y - \bar{\alpha } = \mathrm{tr} B = -p_1 < 0 \) by (8.1), we have
because
Since \(X -\bar{\alpha } < X < - m'(\bar{v})\bar{v} < 0 \) by \(0 < \mu (\bar{u}) < 1\), we have \( d_2^* < d^*\), where \(d^*\) is given by (3.7). Thus, we see that \(J(d_u) > 0\) holds for \( d_u \ge d^*\).
The second assertion of this lemma follows from \( \mathrm{tr} B_{13}= Y - \bar{\alpha } > 0\) and
for \(d_u = 0\). \(\square \)
Remark 6
It should be noted that one cannot skip the proof of the first assertion of Lemma 8.2. For 3-component reaction–diffusion systems, some necessary and sufficient conditions for Turing instability were given by (Cross 1978; Satnoianu et al. 2000; Wang and Li 2001). However, they cannot distinguish the type of Turing instability. On the other hand, (White and Gilligan 1998) gave some necessary conditions for oscillatory Turing instability. To the best of our knowledge, convenient necessary and sufficient conditions for oscillatory Turing instability seem to be unknown.
Rights and permissions
About this article
Cite this article
Kuwamura, M. Turing instabilities in prey–predator systems with dormancy of predators. J. Math. Biol. 71, 125–149 (2015). https://doi.org/10.1007/s00285-014-0816-5
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00285-014-0816-5