Home Mathematics Asymptotic behavior of solutions of a free boundary model with seasonal succession and impulsive harvesting
Article Open Access

Asymptotic behavior of solutions of a free boundary model with seasonal succession and impulsive harvesting

  • Yanglei Li , Xuemei Han and Ningkui Sun EMAIL logo
Published/Copyright: August 13, 2025

Abstract

In this article, we consider the influence of seasonal succession and impulsive harvesting on the dynamical behavior of solutions to a free boundary model. First, the generalized principal eigenvalue is defined and its properties are studied. Next, some sufficient conditions for spreading and for vanishing are given. Then, by introducing a one-parameter family of initial data σ ϕ with σ 0 and ϕ being a compactly supported function, we obtain a threshold value σ * such that spreading happens when σ > σ * , vanishing happens when σ σ * . Finally, we prove the existence and uniqueness of T -periodic traveling semi-wave. Moreover, when spreading persists, we use the proved T -periodic traveling semi-wave to estimate the asymptotic speed of the free boundaries.

MSC 2010: 35K20; 35K57; 35R35; 35B40

1 introduction

The purpose of this article is to consider the combined effects of seasonal succession and impulsive harvesting on the dynamical behavior of solutions to the following problem:

(P) u t = δ u , t ( ( m T ) + , m T + τ ] , x ( g ( t ) , h ( t ) ) , u t = d u x x + a u b u 2 , t ( m T + τ , ( m + 1 ) T ] , x ( g ( t ) , h ( t ) ) , u ( ( m T ) + , x ) = H ( u ( m T , x ) ) , x ( g ( m T ) , h ( m T ) ) , u ( t , g ( t ) ) = 0 = u ( t , h ( t ) ) , t > 0 , g ( t ) = g ( m T ) , h ( t ) = h ( m T ) , t [ m T , m T + τ ] , g ( t ) = μ u x ( t , g ( t ) ) , t ( m T + τ , ( m + 1 ) T ) , h ( t ) = μ u x ( t , h ( t ) ) , t ( m T + τ , ( m + 1 ) T ) , g ( 0 ) = h ( 0 ) = h 0 , u ( 0 , x ) = u 0 ( x ) , x [ h 0 , h 0 ] .

Problem (P) can be used to model the spreading of an invasion species, whose growth and survival are affected by seasonal succession and impulsive harvesting, and its density is stood for u ( t , x ) ; the positive constants T and τ account for the period of seasonal succession and the duration of the bad season, respectively; δ , d , a , and b are positive constants; x = g ( t ) and x = h ( t ) are two moving boundaries representing the expanding fronts of the species, which will be determined together with u ( t , x ) , and μ > 0 describes the ability for spreading into the new habitat (see [3,20] for more background); the initial function u 0 is selected in X ( h 0 ) for some h 0 > 0 , with

(1.1) X ( h 0 ) { ϕ C 2 ( [ h 0 , h 0 ] ) : ϕ ( ± h 0 ) = 0 , ϕ ( x ) > 0 in ( h 0 , h 0 ) } ;

the harvesting function H ( u ) satisfies

  1. H ( u ) C 2 ( [ 0 , ) ) , 1 > H ( 0 ) > 0 = H ( 0 ) , H ( u ) 0 for u > 0 , H ( u ) u is nonincreasing for u , and there are constants D > 0 , η > 1 , and ρ > 0 such that H ( u ) H ( 0 ) u D u η for 0 u ρ .

Here 1 H ( u ) u is used to stand for harvesting rate. There are two classical forms of linear function H ( u ) = c u ( 0 < c < 1 ) and Beverton-Holt function [2]: H ( u ) = β u α + u , where α and β are two constants satisfying 0 < β < α . We set m = 0 , 1 , 2 , , in this article.

In the special case that τ = 0 and H ( u ) = u , that is, there is neither seasonal succession nor impulsive harvesting, then problem (P) becomes

(1.2) u t = d u x x + a u b u 2 , t > 0 , x ( g ( t ) , h ( t ) ) , u ( t , g ( t ) ) = 0 = u ( t , h ( t ) ) , t > 0 , g ( t ) = μ u x ( t , g ( t ) ) , t > 0 , h ( t ) = μ u x ( t , h ( t ) ) , t > 0 , g ( 0 ) = h ( 0 ) = h 0 , u ( 0 , x ) = u 0 ( x ) , x [ h 0 , h 0 ] .

This problem was innovated and studied in [7]. It was shown in [7] that either spreading happens (i.e., h ( t ) , g ( t ) and u ( t , ) a b as t locally uniformly in R ) or vanishing happens (i.e., h ( t ) g ( t ) is bounded and u ( t , ) 0 uniformly as t ), and when spreading happens, the boundaries g ( t ) and h ( t ) behave like k 0 t for large time, where k 0 = k 0 ( μ ) is determined uniquely by the following elliptic problem in half line:

q + k q + a q b q 2 = 0 , q > 0 , z > 0 , q ( ) = a b , q ( 0 ) = 0 , k = μ q ( 0 ) .

They also found that as μ increases to infinity, k 0 ( μ ) increases to 2 a d . The number 2 a d is called the minimal speed of the traveling waves of Fisher-KPP equations (cf. [1]). For further related works, we refer to [5,8,9,17,28,29,31,32] and references therein.

When H ( u ) = u , that is, there is no impulsive harvesting, then problem (P) becomes

(1.3) u t = δ u , t ( m T , m T + τ ] , x ( g ( t ) , h ( t ) ) , u t = d u x x + a u b u 2 , t ( m T + τ , ( m + 1 ) T ] , x ( g ( t ) , h ( t ) ) , u ( t , g ( t ) ) = 0 = u ( t , h ( t ) ) , t > 0 , g ( t ) = g ( m T ) , h ( t ) = h ( m T ) , t [ m T , m T + τ ] , g ( t ) = μ u x ( t , g ( t ) ) , t ( m T + τ , ( m + 1 ) T ) , h ( t ) = μ u x ( t , h ( t ) ) , t ( m T + τ , ( m + 1 ) T ) , g ( 0 ) = h ( 0 ) = h 0 , u ( 0 , x ) = u 0 ( x ) , x [ h 0 , h 0 ] .

Problem (1.3) was introduced and discussed by Peng and Zhao [25]. They examined the effect of the seasonal succession on the dynamical behavior of solutions and found that when a ( T τ ) δ τ 0 , only vanishing happens regardless of the values of h 0 , μ , and u 0 ; while when a ( T τ ) δ τ > 0 , there is a spreading-vanishing dichotomy result. Furthermore, when spreading happens, they used the time periodic traveling semi-waves to estimate the asymptotic spreading speed. For further related works on seasonal succession we refer to [4,10,13,14,18,34,35,40,41].

When τ = 0 and H ( u ) u , i.e., there is no seasonal succession, problem (P) changes into

(1.4) u t = d u x x + a u b u 2 , t ( ( m T ) + , ( m + 1 ) T ] , x ( g ( t ) , h ( t ) ) , u ( ( m T ) + , x ) = H ( u ( m T , x ) ) , x ( g ( m T ) , h ( m T ) ) , u ( t , g ( t ) ) = 0 = u ( t , h ( t ) ) , t > 0 , g ( t ) = μ u x ( t , g ( t ) ) , t > 0 , h ( t ) = μ u x ( t , h ( t ) ) , t > 0 , g ( 0 ) = h ( 0 ) = h 0 , u ( 0 , x ) = u 0 ( x ) , x [ h 0 , h 0 ] .

The above problem was ushered and investigated by Meng et al. [21]. They considered the effect of impulsive harvesting on the dynamical behavior of solutions and showed that if H ( 0 ) e a T , only vanishing happens; while if H ( 0 ) > e a T , either vanishing or spreading happens for a solution which depends on the values of h 0 , μ , and u 0 . The corresponding Cauchy problem for problem (1.4) was introduced by [12,16]. They studied the minimal spreading speed of traveling waves, gave the minimal domain size, and showed how the pulse affects extinction and persistence on species. For further related works on impulsive harvesting, we refer to [11,22,36,38,39] and references therein.

The main purpose of this article is to study the effects of seasonal succession and impulsive harvesting on the dynamical behavior of solutions to problem (P). When 0 < H ( 0 ) e δ τ a ( T τ ) , it follows from [21,25] that u 0 as t . Therefore, throughout this article, we assume that e δ τ a ( T τ ) < H ( 0 ) < 1 . Let us show the existence and uniqueness of solution to problem (P). Firstly, we take m = 0 , then impulsive harvesting takes place at time t = 0 and u ( 0 + , x ) = H ( u 0 ( x ) ) for x [ h 0 , h 0 ] . When t ( 0 + , τ ] , we see that g ( t ) = h ( t ) = h 0 and it follows from the first equation of problem (P) that u ( t , x ) = e δ t H ( u 0 ( x ) ) for ( t , x ) ( 0 + , τ ] × [ h 0 , h 0 ] . When t ( τ , T ] , then ( u , g , h ) satisfies problem (P) with the initial data and initial region at time t = τ replaced by e δ τ H ( u 0 ( x ) ) and [ g ( τ ) , h ( τ ) ] , respectively. It follows from [7,33] that when t ( τ , T ] , problem (P) admits a unique bounded solution ( u , g , h ) . Later, by taking m = 1 , 2 , , recursively, we obtain the existence and uniqueness of solution ( u , g , h ) of problem (P) on [ 0 , ) , which satisfies that

u C 1 , 2 ( ( m T , ( m + 1 ) T ] × [ g ( t ) , h ( t ) ] ) C 1 + ν 2 , 2 + ν ( ( m T + τ , ( m + 1 ) T ] × [ g ( t ) , h ( t ) ] ) , and g , h C ( [ 0 , ) ) C 1 + ν 2 ( [ τ , ) \ { m T + τ } ) for some ν ( 0 , 1 ) .

Moreover, we check that u > 0 in [ 0 , ) × ( g ( t ) , h ( t ) ) , and g ( t ) < 0 < h ( t ) in ( m T + τ , ( m + 1 ) T ] , then g ( t ) and h ( t ) are nondecreasing in t > 0 . Denote

g lim t g ( t ) , h lim t h ( t ) , and I ( g , h ) .

Our first main result is devoted to the asymptotic behavior of solutions to problem (P), which is stated as follows.

Theorem 1.1

Assume that e δ τ a ( T τ ) < H ( 0 ) < 1 and (H) holds. Let ( u , g , h ) be a time-global solution of problem (P) with u 0 = σ ϕ for some ϕ X ( h 0 ) and σ 0 . Then there exists σ * = σ * ( h 0 , ϕ ) [ 0 , ] such that

  1. when σ σ * , vanishing happens in the sense that h g 2 l * and

    lim t u ( t , ) L ( [ g ( t ) , h ( t ) ] ) = 0 ;

  2. when σ > σ * , spreading happens in the sense that I = R and

    lim t [ u ( t , ) P * ( t ) ] = 0 locally uniformly in R ,

where l * and P * ( t ) are given in Lemmas 2.4 and 3.3, respectively.

From the above theorem and its proof, we see that there is a critical value l * for h 0 such that when h 0 l * , σ * = 0 regardless of the choice of ϕ X ( h 0 ) ; when h 0 < l * , σ * ( 0 , ] . Based on the comparison principle, the proof of Theorem 1.1 is given in Section 3.

When spreading happens, it is interesting to study the asymptotic spreading speed. It is seen from Section 4 that the following problem:

(1.5) U t = δ U , t ( 0 + , τ ] , z ( 0 , ) , U t d U z z + k ( t ) U z = a U b U 2 , t ( τ , T ] , z ( 0 , ) , U ( 0 + , z ) = H ( U ( 0 , z ) ) , z ( 0 , ) , U ( t , 0 ) = 0 , t ( 0 , T ) , U ( 0 , z ) = U ( T , z ) , z ( 0 , ) , k ( t ) = μ U z ( t , 0 ) , t ( τ , T ) ,

admits a unique positive solution ( k μ * , U * ) , where k μ * ( t ) is T -periodic in t , k μ * ( t ) 0 in [ 0 , τ ] , and k μ * ( t ) > 0 in ( τ , T ) . Here U * t , 0 t k μ * ( s ) d s x is called a time periodic traveling semi-wave. Thanks to this, we obtain the following result.

Theorem 1.2

Assume that e δ τ a ( T τ ) < H ( 0 ) < 1 , (H) holds and spreading happens for ( u , g , h ) . Let ( k μ * , U * ) be the unique positive solution of problem (1.5). Then

(1.6) lim t g ( t ) t = lim t h ( t ) t = 1 T τ T k μ * ( s ) d s .

The proof of this theorem is given in Section 4, which is divided into two parts: the first part covers the existence and uniqueness of positive solutions to problem (1.5); in the second part we construct some suitable super- and sub-solutions of problem (P) to prove (1.6).

The rest of the article is organized as follows. Section 2 covers some basic results which are useful for this research, and may have other applications. In Section 3, we study the long-time behavior of solutions to problem (P). Section 3.1 is filled with the vanishing phenomenon. Section 3.2 is devoted to the spreading phenomenon. The proof of Theorem 1.1 is given in Section 3.3. In Section 4, we show that problem (1.5) admits a unique positive solution ( k μ * , U * ) , which can be suitably modified to construct super- and sub-solutions of problem (P) to describe the asymptotic spreading speed of spreading solutions as stated in Theorem 1.2. Finally, a brief discussion is given in Section 5.

2 Some basic results

In this section, we give some basic results which will be frequently used later in the article.

2.1 Comparison principle

In this subsection, we establish the comparison principle, which will be used in the rest of this article. Let us start with the following result.

Lemma 2.1

Assume that(H) holds, M ˆ > 0 is an integer, g ˆ , h ˆ C ( [ 0 , M ˆ T ] ) C 1 ( ( m T + τ , ( m + 1 ) T ] ) , ( m = 0 , 1 , 2 , , M ˆ 1 ) and u ˆ ( t , x ) C 1 , 2 ( ( 0 , M ˆ T ] × ( g ˆ , h ˆ ) ) , and

u ˆ t δ u ˆ , t ( ( m T ) + , m T + τ ] , x ( g ˆ ( t ) , h ˆ ( t ) ) , u ˆ t d u ˆ x x + a u ˆ b u ˆ 2 , t ( m T + τ , ( m + 1 ) T ] , x ( g ˆ ( t ) , h ˆ ( t ) ) , u ˆ ( ( m T ) + , x ) H ( u ˆ ( m T , x ) ) , x ( g ˆ ( m T ) , h ˆ ( m T ) ) , u ˆ ( t , g ˆ ( t ) ) = u ˆ ( t , h ˆ ( t ) ) = 0 , t [ 0 , M ˆ T ] , g ˆ ( t ) g ˆ ( m T ) , h ˆ ( t ) h ˆ ( m T ) , t ( m T , m T + τ ] , h ˆ ( t ) μ u ˆ x ( t , h ˆ ( t ) ) , t ( m T + τ , ( m + 1 ) T ] , g ˆ ( t ) μ u ˆ x ( t , g ˆ ( t ) ) , t ( m T + τ , ( m + 1 ) T ] .

If u 0 ( x ) u ˆ ( 0 , x ) for x [ h 0 , h 0 ] [ g ˆ ( 0 ) , h ˆ ( 0 ) ] , and ( u , g , h ) is a solution to problem (P), then

[ g ( t ) , h ( t ) ] [ g ˆ ( t ) , h ˆ ( t ) ] f o r t ( 0 , M ˆ T ] , u ( t , x ) u ˆ ( t , x ) f o r t ( 0 , M ˆ T ] a n d x ( g ( t ) , h ( t ) ) .

Lemma 2.2

Assume that(H) holds, M ˆ > 0 is an integer, g ˆ , h ˆ C ( [ 0 , M ˆ T ] ) C 1 ( ( m T + τ , ( m + 1 ) T ] ) , ( m = 0 , 1 , 2 , , M ˆ 1 ) and u ˆ ( t , x ) C 1 , 2 ( ( 0 , M ˆ T ] × ( g ˆ , h ˆ ) ) , and

u ˆ t δ u ˆ , t ( ( m T ) + , m T + τ ] , x ( g ˆ ( t ) , h ˆ ( t ) ) , u ˆ t d u ˆ x x + a u ˆ b u ˆ 2 , t ( m T + τ , ( m + 1 ) T ] , x ( g ˆ ( t ) , h ˆ ( t ) ) , u ˆ ( ( m T ) + , x ) H ( u ˆ ( m T , x ) ) , x ( g ˆ ( m T ) , h ˆ ( m T ) ) , u ˆ ( t , g ˆ ( t ) ) u ( t , g ˆ ( t ) ) , u ˆ ( t , h ˆ ( t ) ) = 0 , t [ 0 , M ˆ T ] , h ˆ ( t ) h ˆ ( m T ) , t ( m T , m T + τ ] , h ˆ ( t ) μ u ˆ x ( t , h ˆ ( t ) ) , t ( m T + τ , ( m + 1 ) T ] , g ˆ ( t ) g ( t ) , t [ 0 , M ˆ T ] , h 0 h ˆ ( 0 ) , u 0 ( x ) u ˆ ( 0 , x ) , x [ g ˆ ( 0 ) , h 0 ] ,

where ( u , g , h ) is a solution to problem (P). Then

h ( t ) h ˆ ( t ) , u ( t , x ) u ˆ ( t , x ) f o r t ( 0 , M ˆ T ] a n d g ˆ ( t ) < x < h ( t ) .

By a similar argument to that of [7, Lemma 5.7], we can prove Lemma 2.1, and a minor modification of this proof implies Lemma 2.2.

Remark 2.3

The function u ˆ , or the triple ( u ˆ , g ˆ , h ˆ ) , in above lemmas, is known as a supersolution of problem (P). A subsolution can be defined analogously by reversing all the inequalities. The corresponding comparison results for subsolutions in each case can be obtained.

2.2 Linear eigenvalue problem

For any fixed l > 0 , let us consider the following linear eigenvalue problem:

(2.1) φ t = δ φ + λ φ , t ( 0 + , τ ] , x ( l , l ) , φ t = d φ x x + a φ + λ φ , t ( τ , T ] , x ( l , l ) , φ ( 0 + , x ) = H ( 0 ) φ ( 0 , x ) , x ( l , l ) , φ ( 0 , x ) = φ ( T , x ) , x ( l , l ) , φ ( t , ± l ) = 0 , t [ 0 , T ] .

Let us define the generalized principal eigenvalue

λ 1 ( l ) sup { λ R : ( λ , φ ) satisfies (2.2) and φ > 0 in [ 0 , T ] × ( l , l ) } ,

where

(2.2) φ t δ φ + λ φ , t ( 0 + , τ ] , x ( l , l ) , φ t d φ x x + a φ + λ φ , t ( τ , T ] , x ( l , l ) , φ ( 0 + , x ) = H ( 0 ) φ ( 0 , x ) , x ( l , l ) , φ ( 0 , x ) = φ ( T , x ) , x ( l , l ) , φ ( t , ± l ) = 0 , t [ 0 , T ] .

We obtain the following result on the properties of the generalized principal eigenvalue of problem (2.1).

Lemma 2.4

For any fixed l > 0 , let λ 1 ( l ) be the generalized principal eigenvalue of problem (2.1), then

(2.3) λ 1 ( l ) = d π 2 4 l 2 T τ T + δ τ T a T τ T 1 T ln H ( 0 ) .

  1. If 0 < H ( 0 ) e δ τ a ( T τ ) , λ 1 ( l ) > 0 for all l > 0 .

  2. If e δ τ a ( T τ ) < H ( 0 ) < 1 , there is a positive constant

    (2.4) l * 1 2 d π 2 ( T τ ) a ( T τ ) δ τ + ln H ( 0 )

    such that λ 1 ( l ) is negative (resp. 0, or positive) when l > l * (resp. l = l * , or l < l * ).

Proof

First, let us show that λ 1 ( l ) given in (2.3) satisfies problem (2.1). Indeed, set

(2.5) φ * ( t , x ) = cos π x 2 l ζ ( t ) ,

where ζ ( t ) satisfies

(2.6) ζ ( t ) 1 H ( 0 ) , t = 0 , e ( λ 1 ( l ) δ ) t , t ( 0 + , τ ] , e λ 1 ( l ) t + a d π 2 4 l 2 ( t τ ) δ τ , t ( τ , T ] ,

we can check that ( λ 1 ( l ) , φ * ( t , x ) ) satisfies problem (2.1), where φ * ( t , x ) is given by (2.5).

Next we prove that λ 1 ( l ) is the generalized principal eigenvalue, that is, if ( λ ˜ , φ ˜ ) satisfies problem (2.2), then λ ˜ is not greater than λ 1 ( l ) . In fact, set φ ˆ ( t , x ) = ζ 1 ( t ) cos π x 2 l , we can check that

(2.7) φ ˆ t = δ φ ˆ + λ 1 ( l ) φ ˆ , t ( 0 + , τ ] , x ( l , l ) , φ ˆ t = d φ ˆ x x + a φ ˆ + λ 1 ( l ) φ ˆ , t ( τ , T ] , x ( l , l ) , φ ˆ ( 0 + , x ) = 1 H ( 0 ) φ ˆ ( 0 , x ) , x ( l , l ) , φ ˆ ( 0 , x ) = φ ˆ ( T , x ) , x ( l , l ) , φ ˆ ( t , ± l ) = 0 , t [ 0 , T ] .

Multiplying the first two inequalities in problem (2.2) with ( λ , φ ) replaced with ( λ ˜ , φ ˜ ) by φ ˆ , multiplying the first two equations in problem (2.7) by φ ˜ , subtracting the two results and integrating over ( 0 + , T ] × [ l , l ] , we obtain that

(2.8) l l 0 + T ( φ ˜ t φ ˆ + φ ˆ t φ ˜ ) d t d x d l l τ T ( φ ˜ x x φ ˆ φ ˆ x x φ ˜ ) d t d x l l 0 + T ( λ ˜ λ 1 ( l ) ) φ ˜ φ ˆ d t d x .

It follows from the divergence formula, the boundary condition, and the periodic condition that

l l 0 + T ( φ ˜ t φ ˆ + φ ˆ t φ ˜ ) d t d x d l l τ T ( φ ˜ x x φ ˆ φ ˆ x x φ ˜ ) d t d x = l l [ φ ˜ ( T , x ) φ ˆ ( T , x ) φ ˜ ( 0 + , x ) φ ˆ ( 0 + , x ) ] d x d τ T ( φ ˜ x φ ˆ φ ˆ x φ ˜ ) l l d t = l l [ φ ˜ ( T , x ) φ ˆ ( T , x ) φ ˜ ( 0 , x ) φ ˆ ( 0 , x ) ] d x = 0 .

This, together with (2.8) and the fact that φ ˜ > 0 and φ ˆ > 0 in [ 0 , T ] × ( l , l ) , implies that λ ˜ λ 1 ( l ) , thus the generalized principal eigenvalue of problem (2.2) is given by (2.3). Moreover, we can check that λ 1 ( l ) is strictly decreasing in l > 0 .

(i) If 0 < H ( 0 ) e δ τ a ( T τ ) , it follows that

λ 1 ( l ) d π 2 4 l 2 T τ T > 0 for all l > 0 .

(ii) If e δ τ a ( T τ ) < H ( 0 ) < 1 , we can check that λ 1 ( l * ) = 0 . Combining this with the monotonicity of λ 1 ( l ) in l , we see that the other assertions hold.□

3 Classification of the long-time behavior and sharp thresholds

The purpose of this section is to study the long-time behavior of solutions. The first subsection covers the vanishing phenomenon. The second subsection deals with the spreading phenomenon. The proof of Theorem 1.1 is given in the last subsection.

3.1 Vanishing phenomenon

This subsection starts with the following result.

Lemma 3.1

Assume that e δ τ a ( T τ ) < H ( 0 ) < 1 and (H) holds. Let ( u , g , h ) be a solution of problem (P) and l * be given in (2.4). Then the following results are equivalent:

  1. h or g is finite;

  2. h g 2 l * ;

  3. u ( t , ) L ( [ g ( t ) , h ( t ) ] ) 0 a s t .

Proof

“(i) (ii)”. Without loss of generality we assume that g > . If h g > 2 l * , there exists an integer n 1 1 such that h ( t 1 ) g ( t 1 ) > 2 l * for t 1 n 1 T . To obtain a contradiction, let us consider the following problem:

(3.1) v t = δ v , t ( ( m ˆ T ) + , m ˆ T + τ ] , x ( ( t ) , h ( t 1 ) ) , v t = d v x x + a v b v 2 , t ( m ˆ T + τ , ( m ˆ + 1 ) T ] , x ( ( t ) , h ( t 1 ) ) , v ( ( m ˆ T ) + , x ) = H ( v ( m ˆ T , x ) ) , x ( ( m ˆ T ) , h ( t 1 ) ) , ( t ) = ( m ˆ T ) , t [ m ˆ T , m ˆ T + τ ] , ( t ) = μ v x ( t , ( t ) ) , t ( m ˆ T + τ , ( m ˆ + 1 ) T ] , v ( t , h ( t 1 ) ) = v ( t , ( t ) ) = 0 , t t 1 , ( t 1 ) = g ( t 1 ) , v ( t 1 , x ) = u ( t 1 , x ) , x [ g ( t 1 ) , h ( t 1 ) ] ,

where m ˆ = n 1 , n 1 + 1 , n 1 + 2 , . It is direct to check that v is a subsolution of problem (P) and ( t ) g ( t ) for t t 1 . Note that we have assumed that g > , then ( ) > . It follows from a similar argument as in the proof of [6, Lemma 3.3] that

v ( t , ) V ( t , ) C 2 ( [ ( t ) , h ( t 1 ) ] ) 0 as t ,

where V ( t , x ) is the unique T -periodic solution of

(3.2) V t = δ V , t ( 0 + , τ ] , x ( ( ) , h ( t 1 ) ) , V t = d V x x + a V b V 2 , t ( τ , T ] , x ( ( ) , h ( t 1 ) ) , V ( 0 + , x ) = H ( V ( 0 , x ) ) , x ( ( ) , h ( t 1 ) ) , V ( t , ( ) ) = V ( t , h ( t 1 ) ) = 0 , t [ 0 , T ] , V ( 0 , x ) = V ( T , x ) , x ( ( ) , h ( t 1 ) ) .

Recalling that h ( t 1 ) ( ) h ( t 1 ) g ( t 1 ) > 2 l * , then we see that V > 0 in [ 0 , T ] × ( ( ) , h ( t 1 ) ) . As n , we have

( s + n T ) = μ v x ( s + n T , ( s + n T ) ) μ V x ( s , ( ) ) < β uniformly on T + τ 2 , 2 T + τ 3 ,

for some β > 0 . Combining this with the monotonicity of ( t ) in t > 0 , we see that ( ) = , which is a contradiction. Thus, we obtain that h g 2 l * .

“(ii) (i)”. When (ii) holds, (i) is obvious.

“(ii) (iii)”. As h g 2 l * , it then follows from [22] that the unique positive solution of the following problem:

(3.3) w t = δ w , t ( ( m T ) + , m T + τ ] , x ( g , h ) , w t = d w x x + a w b w 2 , t ( m T + τ , ( m + 1 ) T ] , x ( g , h ) , w ( ( m T ) + , x ) = H ( w ( m T , x ) ) , x ( g , h ) , w ( t , g ) = w ( t , h ) = 0 , t > 0 , w ( 0 , x ) = w 0 ( x ) , 0 , x [ g , h ] ,

with w 0 ( x ) u 0 ( x ) for x [ h 0 , h 0 ] , satisfies w 0 uniformly for x [ g , h ] as t . Then lim t u ( t , ) L ( [ g ( t ) , h ( t ) ] ) = 0 follows easily from the comparison principle.

“(iii) (ii)”: Assume that, for some small ε > 0 there is a large integer n 2 such that

h ( t ) g ( t ) > 2 l * + 3 ε for t > t 2 n 2 T .

Thanks to Lemma 2.4, we know that problem (2.1) with l = l * + ε admits a generalized principal eigenvalue λ ε , which satisfies that λ ε < 0 , whose corresponding positive eigenfunction φ ε can be normalized by φ ε L = 1 . Set

u ̲ ( t , x ) γ φ ε ( t , x ) in [ 0 , T ] × Ω * ,

where Ω * ( l * ε , l * + ε ) and γ > 0 is small such that λ ε > 4 b γ . It is direct to check that

(3.4) u ̲ t + δ u ̲ = λ ε u ̲ 0 , t ( ( m T ) + , m T + τ ] , x Ω * , u ̲ t d u ̲ x x a u ̲ + b u ̲ 2 = λ ε u ̲ + b u ̲ 2 1 2 λ ε u ̲ 0 , t ( m T + τ , ( m + 1 ) T ] , x Ω * , u ̲ ( ( m T ) + , x ) = H ( u ̲ ( m T , x ) ) , x Ω * , u ̲ ( t , l * ε ) = u ̲ ( t , l * + ε ) = 0 , t > 0 .

We can also choose γ small such that

γ φ ε ( 0 , x ) < u t 2 , x h ( t 2 ) g ( t 2 ) 2 for x Ω * .

The comparison principle can be used to show that

u ( t + t 2 , g ( t 2 ) h ( t 2 ) 2 ) u ̲ ( t , 0 ) > 0 for t 0 ,

which is in contradiction to assertion (iii). The proof of this lemma is complete.□

Next, a sufficient condition for vanishing is given as follows.

Lemma 3.2

Assume that e δ τ a ( T τ ) < H ( 0 ) < 1 and (H) holds. Let ( u , g , h ) be a solution of problem (P) and l * be given in (2.4). If h 0 < l * and if u 0 L ( [ h 0 , h 0 ] ) is sufficiently small, then vanishing happens.

Proof

For any fixed h 1 ( h 0 , l * ) , it is seen from Lemma 2.4 and [21] that problem (2.1) with l = h 1 admits a generalized principal eigenvalue λ 1 , which satisfies that λ 1 > 0 , whose corresponding positive eigenfunction can be chosen as ζ ( t ) cos π 2 h 1 x , where ζ ( t ) is a positive solution to the following problem:

ζ t = ( λ 1 δ ) ζ , t ( 0 + , τ ] , ζ t = λ 1 + a d π 2 4 h 1 2 ζ , t ( τ , T ] , ζ ( 0 + ) = H ( 0 ) ζ ( 0 ) , ζ ( 0 ) = ζ ( T ) , ζ ( t ) 1 , t [ 0 , T ] .

Choose

α min { λ 1 , h 1 h 0 1 } ,

then we can find ε = ε ( α ) > 0 small such that

π μ ε < α 2 h 0 2 .

Define

r ( t ) h 0 1 + α α 2 e α t for t > 0 , w ( t , x ) ε e α t ζ ( t ) cos π x 2 r ( t ) for t > 0 , x [ r ( t ) , r ( t ) ] .

A direct calculation shows that

w t + δ w ( λ 1 α ) w 0 , t ( ( m T ) + , m T + τ ] , x ( r ( t ) , r ( t ) ) , w t d w x x a w + b w 2 λ 1 d π 2 4 h 1 2 α + d π 2 4 h 0 2 ( 1 + α ) 2 w 0 , t ( m T + τ , ( m + 1 ) T ] , x ( r ( t ) , r ( t ) ) , w ( ( m T ) + , x ) = H ( 0 ) w ( m T , x ) H ( w ( m T , x ) ) , x ( r ( m T ) , r ( m T ) ) , r ( t ) r ( m T ) , t ( m T , m T + τ ] , w ( t , ± r ( t ) ) = 0 , t > 0 .

Moreover, we can compute that for t ( m T + τ , ( m + 1 ) T ] ,

μ w x ( t , r ( t ) ) = μ w x ( t , r ( t ) ) = π μ ε ζ ( t ) 2 r ( t ) e α t π μ ε 2 h 0 e α t α 2 h 0 2 e α t = r ( t ) .

As h 0 < r ( 0 ) = h 0 ( 1 + α 2 ) , if we choose u 0 small satisfying

u 0 L ( [ h 0 , h 0 ] ) ε ζ ( 0 ) cos π 2 + α ,

then

u 0 ( x ) w ( 0 , x ) in [ h 0 , h 0 ] ,

it thus follows from Lemma 2.1 that h ( t ) r ( t ) for t > 0 , which produces that

h < h 0 ( 1 + α ) .

This, together with Lemma 3.1, implies that vanishing happens. The proof is complete.□

3.2 Spreading phenomenon

The purpose of this subsection is to study the spreading phenomenon. Let us start with the following result.

Lemma 3.3

Assume that e δ τ a ( T τ ) < H ( 0 ) < 1 and (H) holds. Let ( u , g , h ) be a solution of problem (P). If h = g = , then

(3.5) lim t [ u ( t , ) P * ( t ) ] = 0 l o c a l l y u n i f o r m l y i n R ,

where P * ( t ) is the unique positive solution of the problem

(3.6) P t = δ P , t ( 0 + , τ ] , P t = a P b P 2 , t ( τ , T ] , P ( 0 + ) = H ( P ( 0 ) ) , P ( 0 ) = P ( T ) .

Proof

First, we show the existence and uniqueness of the positive solution P * ( t ) to problem (3.6). Motivated by [22], it follows from the first two equations in problem (3.6) that

P ( T ) = a H ( P ( 0 ) ) e δ τ a e a ( T τ ) + b [ 1 e a ( T τ ) ] H ( P ( 0 ) ) e δ τ .

As P ( 0 ) = P ( T ) , if we define

(3.7) F ( s ) a B s C s H ( s ) for s 0 ,

where

B b [ 1 e a ( T τ ) ] > 0 and C a e δ τ a ( T τ ) > 0 .

It is direct to check that problem (3.6) admits a positive solution if and only if F ( s ) = 0 has a positive root. Recalling that H ( 0 ) > e δ τ a ( T τ ) , we have

lim s 0 + F ( s ) = a H ( 0 ) [ H ( 0 ) e δ τ a ( T τ ) ] > 0 .

Since H ( s ) 0 for s 0 , and B , C > 0 , we can check that

F ( s ) a B s < 0 for s > a B .

Thus, the existence of positive solution to problem (3.6) is proved. Combining with the monotonicity of H ( s ) s in s 0 , we see that F ( s ) is strictly decreasing in s 0 , so the uniqueness of solution to problem (3.6) follows.

Next, since h = g = , there are l > l * and an integer N l > 0 such that h ( t ) g ( t ) 2 l for t N l T , where l * is given in (2.4). Let v ̲ ( t , x ) be the unique positive solution of

(3.8) v ̲ t = δ v ̲ , t ( ( m ˆ T ) + , m ˆ T + τ ] , x ( l , l ) , v ̲ t = d v ̲ x x + a v ̲ b v ̲ 2 , t ( m ˆ T + τ , ( m ˆ + 1 ) T ] , x ( l , l ) , v ̲ ( ( m ˆ T ) + , x ) = H ( v ̲ ( m ˆ T , x ) ) , x ( l , l ) , v ̲ ( t , ± l ) = 0 , t N l T , v ̲ ( N l T , x ) = u ( N l T , x ) , x [ l , l ] ,

where m ˆ = N l , N l + 1 , . It follows from the comparison principle that

(3.9) u ( t , x ) v ̲ ( t , x ) for t N l T and x [ l , l ] .

Since l > l * , it then follows from a similar argument as in [24] that the following problem:

(3.10) V t = δ V , t ( 0 + , τ ] , x ( l , l ) , V t = d V x x + a V b V 2 , t ( τ , T ] , x ( l , l ) , V ( 0 + , x ) = H ( V ( 0 , x ) ) , x ( l , l ) , V ( t , ± l ) = 0 , t [ 0 , T ] , V ( 0 , x ) = V ( T , x ) , x [ l , l ] ,

admits a unique positive solution V ̲ ( t , x ; l ) and

(3.11) lim n [ v ̲ ( t + n T , x ) V ̲ ( t , x ; l ) ] = 0 uniformly in [ 0 , T ] × [ l , l ] .

Moreover, as  l , V ̲ ( t , x ; l ) monotonically converges to P * ( t ) locally uniformly in [ 0 , T ] × R . Combining with (3.9) and (3.11), we obtain that

(3.12) liminf n u ( t + n T , x ) P * ( t ) locally uniformly in [ 0 , T ] × R .

Later, it follows from the proof of [13, Lemma 2.1] that the unique solution v ¯ ( t ) of

(3.13) v ¯ t = δ v ¯ , t ( ( m T ) + , m T + τ ] , v ¯ t = a v ¯ b v ¯ 2 , t ( m T + τ , ( m + 1 ) T ] , v ¯ ( ( m T ) + ) = H ( v ¯ ( m T ) ) , v ¯ ( 0 ) = u 0 + a b ,

satisfies v ¯ ( t ) P * ( t ) as t . This, together with the comparison principle, yields that

(3.14) limsup n u ( t + n T , x ) P * ( t ) locally uniformly in [ 0 , T ] × R .

This, together with (3.12) and the periodicity of P * ( t ) , produces that (3.5) holds.□

Later, we show the following sufficient condition for spreading.

Lemma 3.4

Assume that e δ τ a ( T τ ) < H ( 0 ) < 1 and (H) holds. Let ( u , g , h ) be a solution of problem (P) and l * be given in (2.4). If h 0 l * , then spreading occurs.

Proof

Noting that g ( t ) < 0 < h ( t ) in ( m T + τ , ( m + 1 ) T ) , then we see that h ( t ) g ( t ) > 2 l * for any t > τ . By Lemmas 3.1 and 3.3, we can check that spreading happens in this case.□

3.3 Proof of Theorem 1.1

By Lemma 3.1, we see that either h g 2 l * or h = g = . This, together with Lemmas 3.1 and 3.3, yields that there is a spreading-vanishing dichotomy result.

Set

σ * sup { σ 0 0 : vanishing happens for σ ( 0 , σ 0 ] } .

When h 0 l * , by Lemma 3.4, we see that σ * = 0 . When h 0 < l * , it follows from Lemma 3.2 that vanishing happens for all small σ > 0 , so σ * ( 0 , + ] . If σ * = , there is nothing left to do. In what follows, we suppose σ * ( 0 , ) . The comparison principle implies that vanishing happens for σ [ 0 , σ * ) . We claim that vanishing happens for σ = σ * . Otherwise, by the proved spreading-vanishing dichotomy result, we find that spreading happens for σ = σ * , which implies that there exists a large integer n 0 > 0 such that h ( t 0 ) g ( t 0 ) > 2 l * + 1 with t 0 n 0 T . By the continuous dependence of the solution of problem (P) on its initial values, we see that if ε > 0 is sufficiently small, the solution ( u ε , g ε , h ε ) of problem (P) with u 0 = ( σ * ε ) ϕ satisfies

h ε ( t 0 ) g ε ( t 0 ) > 2 l * .

This, together with Lemma 3.4, produces that spreading happens for σ * ε , which is in contradiction to the definition of σ * , so vanishing happens for σ = σ * . Combining this with the proved spreading-vanishing dichotomy result and the comparison principle, we show that spreading happens for σ > σ * , which ends the proof.

4 Asymptotic spreading speeds

Throughout this section we assume that e δ τ a ( T τ ) < H ( 0 ) < 1 , (H) holds and ( u , g , h ) is a solution of problem (P) for which spreading happens. The purpose of this section is to investigate the asymptotic spreading speed of free boundaries.

4.1 T -periodic traveling semi-waves

This subsection is devoted to the existence and uniqueness of T -periodic traveling semi-wave. Define

K { p ( t ) C ν 2 ( [ 0 , T ] \ { τ } ) : p ( t ) is T -periodic , p ( t ) 0 in [ 0 , τ ] } ; for each p K and c R , set p c ¯ 1 T τ τ T ( p ( t ) c ) d t = p ¯ c .

For any nonnegative k ( t ) K , consider the following problem:

(4.1) U t = δ U , t ( 0 + , τ ] , z ( 0 , ) , U t d U z z + k ( t ) U z = a U b U 2 , t ( τ , T ] , z ( 0 , ) , U ( 0 + , z ) = H ( U ( 0 , z ) ) , z ( 0 , ) , U ( t , 0 ) = 0 , t [ 0 , T ] , U ( 0 , z ) = U ( T , z ) , z ( 0 , ) .

We obtain the following result.

Theorem 4.1

Assume that e δ τ a ( T τ ) < H ( 0 ) < 1 and (H) holds. For any nonnegative k K , problem (4.1) has a unique positive solution U ( t , z ; k ) if and only if k ¯ [ 0 , c 0 ) where

c 0 2 d a δ τ T τ + ln H ( 0 ) T τ .

Moreover, the unique positive solution U ( t , z ; k ) is continuous in k , U ( t , ; k ) = P * ( t ) for t [ 0 , T ] , and U z ( t , z ; k ) > 0 for ( t , z ) [ 0 , T ] × [ 0 , ) . Furthermore, for any nonnegative function k 1 , k 2 K with k i ¯ [ 0 , c 0 ) ( i = 1 , 2 ) , if k 1 , k 2 , we see that

U ( t , z ; k 1 ) > U ( t , z ; k 2 ) for ( t , z ) [ 0 , T ] × [ 0 , ) .

Proof

The proof is similar to that in [6] (see also [30]), we give the details for the sake of completeness. Our argument is divided into five steps.

Step 1. Existence of a positive and increasing solution when nonnegative k K satisfies k ¯ [ 0 , c 0 ) .

Let us consider the following linear eigenvalue problem:

(4.2) ψ t + δ ψ = λ ψ , t ( 0 + , τ ] , z ( n , n ) , ψ t d ψ z z + k ( t ) ψ z a ψ = λ ψ , t ( τ , T ] , z ( n , n ) , ψ ( 0 + , z ) = H ( 0 ) ψ ( 0 , z ) , z ( n , n ) , ψ ( t , ± n ) = 0 , t [ 0 , T ] , ψ ( 0 , z ) = ψ ( T , z ) , z ( n , n ) .

It follows from [6,21,23] and Lemma 2.4 that problem (4.2) admits a generalized principal eigenvalue λ 1 n , which is given by

λ 1 n sup { λ n R : ( λ n , ψ ) satisfies (4.3) and ψ > 0 in [ 0 , T ] × ( n , n ) } ,

where

(4.3) ψ t + δ ψ λ n ψ , t ( 0 + , τ ] , z ( n , n ) , ψ t d ψ z z + k ( t ) ψ z a ψ λ n ψ , t ( τ , T ] , z ( n , n ) , ψ ( 0 + , z ) = H ( 0 ) ψ ( 0 , z ) , z ( n , n ) , ψ ( t , ± n ) = 0 , t [ 0 , T ] , ψ ( 0 , z ) = ψ ( T , z ) , z ( n , n ) .

We can check that λ 1 n is decreasing in n and λ 1 n λ 1 as n , where λ 1 is the generalized principal eigenvalue of

(4.4) Ψ t + δ Ψ = λ Ψ , t ( 0 + , τ ] , z R , Ψ t d Ψ z z + k ( t ) Ψ z a Ψ = λ Ψ , t ( τ , T ] , z R , Ψ ( 0 + , z ) = H ( 0 ) Ψ ( 0 , z ) , z R , Ψ ( 0 , z ) = Ψ ( T , z ) , z R .

In fact, λ 1 is given by

λ 1 sup { λ R : ( λ , Ψ ) satisfies (4.5) and Ψ > 0 in [ 0 , T ] × R } ,

where

(4.5) Ψ t + δ Ψ λ Ψ , t ( 0 + , τ ] , z R , Ψ t d Ψ z z + k ( t ) Ψ z a Ψ λ Ψ , t ( τ , T ] , z R , Ψ ( 0 + , z ) = H ( 0 ) Ψ ( 0 , z ) , z R , Ψ ( 0 , z ) = Ψ ( T , z ) , z R .

Moreover, by a similar argument as that in Lemma 2.4, we can obtain that

λ 1 = ( T τ ) k ¯ 2 4 d T a T τ T + δ τ T ln H ( 0 ) T .

When k K with k ¯ [ 0 , c 0 ) , it follows that λ 1 < 0 , which implies that for any large n , the generalized principal eigenvalue λ 1 n of problem (4.2) is negative, whose corresponding positive eigenfunction ψ 1 can be normalized by ψ 1 L = 1 . Let us choose ε > 0 sufficiently small satisfying ε < C * a b + 1 and define

U ̲ ( t , z ) = ε ψ 1 ( t , z n ) , ( t , z ) [ 0 , T ] × [ 0,2 n ] , 0 , ( t , z ) [ 0 , T ] × ( 2 n , ) .

Then we can check that U ̲ ( t , z ) is a subsolution of problem (4.1) and U ¯ ( t , z ) C * is a supersolution of problem (4.1). The sub- and supersolution argument and the strong maximum principle can be used to obtain that problem (4.1) admits at least one positive solution U ( t , z ; k ) satisfying U > 0 in [ 0 , T ] × ( 0 , ) . Moreover, if U is a positive solution of problem (4.1), it then follows from [6] that as z , U z ( t , z ; k ) P * ( t ) uniformly in t [ 0 , T ] and

U z ( t , z ; k ) > 0 for ( t , z ) [ 0 , T ] × [ 0 , ) .

Step 2. Non-existence when k K with k ¯ c 0 .

Assume that k K with k ¯ c 0 , then the principal eigenvalue λ 1 of problem (4.4) is nonnegative, whose corresponding positive eigenfunction Ψ 1 ( t , z ) can be selected as

Ψ 1 ( t , z ) e k ¯ 2 d z θ ( t ) for t [ 0 , T ] , z R ,

where θ ( t ) satisfies

θ t + δ θ = λ 1 θ , t ( 0 + , τ ] , θ t + k ¯ 2 d k ( t ) k ¯ 2 4 d a θ = λ 1 θ , t ( τ , T ] , θ ( 0 + ) = H ( 0 ) θ ( 0 ) , θ ( 0 ) = θ ( T ) .

Then Ψ 1 ( t , z ) > 0 for ( t , z ) [ 0 , T ] × [ 0 , ) and

(4.6) Ψ 1 ( t , z ) uniformly on [ 0 , T ] as z .

Suppose by way of contradiction that problem (4.1) admits a positive solution U . It follows from Step 1 that U ( t , z ) C * for ( t , z ) [ 0 , T ] × [ 0 , ) . Let us define

Σ = { ϱ > 0 : ϱ Ψ 1 ( t , z ) U ( t , z ) for ( t , z ) [ 0 , T ] × [ 0 , ) } .

Clearly Σ , which is relatively closed in ( 0 , ) . We shall show that Σ is open. Indeed, if ϱ * Σ , we see that ϱ * Ψ 1 U in [ 0 , T ] × [ 0 , ) . It follows from (4.6) that we can find z 0 > 0 large and ε 1 > 0 small such that

( ϱ * ε ) Ψ 1 ( t , z ) > U ( t , z ) for ( t , z ) [ 0 , T ] × [ z 0 , ) and ε ( 0 , ε 1 ] .

Set W ( t , z ) ϱ * Ψ 1 ( t , z ) U ( t , z ) , it is direct to check that

W t δ W , t ( 0 + , τ ] , z ( 0 , z 0 ) , W t d W z z k ( t ) W z + [ a b ( ϱ * Ψ 1 + U ) ] W , t ( τ , T ] , z [ 0 , z 0 ) , W ( 0 + , z ) H ( ϱ * Ψ 1 ( 0 , z ) ) H ( U ( 0 , z ) ) 0 , z [ 0 , z 0 ) , W ( t , 0 ) > 0 , W ( t , z 0 ) > 0 , t [ 0 , T ] , W ( 0 , z ) = W ( T , z ) , z [ 0 , z 0 ) ,

where the monotonicity of H ( s ) and H ( s ) s in s 0 is used. The strong maximum principle can be used to conclude that there is a positive constant ε 2 ε 1 such that W ε 2 Ψ 1 in [ 0 , T ] × [ 0 , z 0 ) . As a consequence, we have that

( ϱ * ε ) Ψ 1 ( t , z ) U ( t , z ) in [ 0 , T ] × [ 0 , ) for all ε ( 0 , ε 2 ] ,

which yields that Σ is an open subset of ( 0 , ) . Thus Σ = ( 0 , ) , which means that ϱ Ψ 1 ( t , z ) U ( t , z ) in [ 0 , T ] × [ 0 , ) for any ϱ ( 0 , ) . This is in contradiction to the fact that U is positive, which ends the proof of non-existence.

Step 3. Uniqueness when k K is nonnegative and satisfies k ¯ [ 0 , c 0 ) .

Let us fix k K with k ¯ [ 0 , c 0 ) . Suppose that problem (4.1) admits two positive solutions U 1 and U 2 . It follows from the conclusions of Step 1 and the Hopf boundary lemma that U z i ( t , z ) > 0 , U i ( t , ) = P * ( t ) , and U i ( t , 0 ) = 0 for t [ 0 , T ] , z 0 and i = 1 , 2 . Set

ρ * inf { ρ 1 : ρ U 1 ( t , z ) U 2 ( t , z ) for ( t , z ) [ 0 , T ] × [ 0 , ) } .

We claim that ρ * = 1 . Otherwise, ρ * > 1 . Define V ( t , z ) ρ * U 1 ( t , z ) U 2 ( t , z ) . Then V ( t , z ) 0 , V ( t , 0 ) = 0 , and V ( t , ) = ( ρ * 1 ) P * ( t ) > 0 for t [ 0 , T ] , z 0 . A direct calculation shows that

V t = δ V , t ( 0 + , τ ] , z 0 , V t d V z z k ( t ) V z + ( a 2 ρ * b U 1 ) V , t ( τ , T ] , z 0 , V ( 0 + , z ) H ( ρ * U 1 ( 0 , z ) ) H ( U 2 ( 0 , z ) ) 0 , z 0 , V ( 0 , z ) = V ( T , z ) , z 0 ,

where ρ * > 1 and the monotonicity of H ( s ) and H ( s ) s in s 0 are used. Thanks to the monotonicity of H ( s ) s in s 0 and 0 < H ( 0 ) < 1 , it then follows from the strong maximum principle and Hopf’s boundary lemma that V κ U 2 in [ 0 , T ] × [ 0 , ) for some κ > 0 small, which infers that

ρ * 1 + κ U 1 ( t , z ) U 2 ( t , z ) for ( t , z ) [ 0 , T ] × [ 0 , ) .

This is in contradiction to the definition of ρ * . Hence, ρ * = 1 and U 1 U 2 in [ 0 , T ] × [ 0 , ) . Similarly, U 2 U 1 in [ 0 , T ] × [ 0 , ) . Thus, U 1 ( t , z ) U 2 ( t , z ) for ( t , z ) [ 0 , T ] × [ 0 , ) , and the desired result follows.

Step 4. Monotonicity of positive solution U ( t , z ; k ) in nonnegative k K with k ¯ [ 0 , c 0 ) .

For any given two functions k i K for i = 1 , 2 satisfying 0 k 1 , k 2 and k 2 ¯ [ 0 , c 0 ) , let U ( t , z ; k i ) be the unique positive solution of problem (4.1) with k = k i . Define

M * inf { M 1 : M U ( t , z ; k 1 ) > U ( t , z ; k 2 ) , ( t , z ) [ 0 , T ] × ( 0 , ) } .

We claim that M * = 1 . Otherwise, M * > 1 . Define V ( t , z ) M * U ( t , z ; k 1 ) U ( t , z ; k 2 ) for ( t , z ) [ 0 , T ] × ( 0 , ) . Then V ( t , z ) 0 , V ( t , 0 ) = 0 , and V ( t , ) = ( M * 1 ) P * ( t ) > 0 for t [ 0 , T ] , z 0 . It follows from a similar argument as in Step 3 that there is a contradiction with the definition of M * . Thus M * = 1 , which means that U ( t , z ; k 1 ) U ( t , z ; k 2 ) in [ 0 , T ] × [ 0 , ) . This, together with the uniqueness of positive solution to problem (4.1), the strong maximum principle and the Hopf boundary lemma, implies that

U ( t , z ; k 1 ) > U ( t , z ; k 2 ) and U z ( t , 0 ; k 1 ) > U z ( t , 0 ; k 2 ) for t [ 0 , T ] , z > 0 ,

which implies the monotonicity of U ( t , z ; k ) in nonnegative k K with k ¯ [ 0 , c 0 ) .

Step 5. Continuity of positive solution U ( t , z ; k ) in nonnegative k K with k ¯ [ 0 , c 0 ) .

For any nonnegative { k n } K with k n k in C ν 2 ( ( τ , T ) ) as n , we shall prove that

(4.7) U ( t , z ; k n ) U ( t , z ; k ) locally in C 1 + ν 2,2 + ν ( ( τ , T ] × [ 0 , ) ) as n .

In fact, thanks to Steps 1–3, we see that when k n K with k n ¯ c 0 , then problem (4.1) with k replaced by k n only admits trivial solution 0; when k n K with 0 k n ¯ < c 0 , then U ( t , z ; k n ) , the unique positive solution of problem (4.1) with k replaced by k n , satisfies 0 < U ( t , z ; k n ) P * ( t ) for t [ 0 , T ] , z > 0 . Moreover, by the standard regularity theory for parabolic equation (up to the boundary), we can find a subsequence of { U ( t , z ; k n ) } (still denoted them by { U ( t , z ; k n ) } such that

U ( t , z ; k n ) U ˜ ( t , z ) locally uniformly in C 1 , 2 ( ( 0 , T ] × [ 0 , ) ) as n ,

where U ˜ is a nonnegative solution of problem (4.1). We next claim that U ˜ ( t , z ) U ( t , z ; k ) in [ 0 , T ] × [ 0 , ) . Indeed, if U ˜ > 0 in [ 0 , T ] × ( 0 , ) , the uniqueness of positive solution to problem (4.1) can be used to show that our claim is true. If U ˜ 0 , we need to show U ( t , z ; k ) 0 in this case and so U ˜ ( t , z ) U ( t , z ; k ) in [ 0 , T ] × [ 0 , ) . Suppose by way of contradiction that U ( t , z ; k ) is a positive solution of problem (4.1). By the above arguments, we see that k K with k ¯ [ 0 , c 0 ) and there is a constant ε > 0 small such that if k ε is an element in the function class K with k ε ( t ) = k ( t ) + ε for t ( τ , T ) , we then see that

k ε ¯ = k ¯ + ε [ 0 , c 0 ) .

Thanks to Steps 1 and 3, we can check that problem (4.1) with k replaced by k ε admits a unique positive solution U ( t , z ; k ε ) . On the other hand, since k n k uniformly on ( τ , T ) , there exists a large integer n ε > 0 such that k n ( t ) k ε ( t ) in [ 0 , T ] for all n n ε . Combining with Steps 1 and 3, we see that for each n n ε , the unique positive solution U ( t , z ; k n ) of problem (4.1) with k replaced by k n satisfies

U ( t , z ; k n ) U ( t , z ; k ε ) in [ 0 , T ] × [ 0 , ) for all n n ε ,

which is in contradiction to the fact that U ( t , z ; k n ) 0 as n locally in C 1 , 2 ( [ 0 , T ] × [ 0 , ) ) . Thus, we see that

U ( t , z ; k n ) U ( t , z ; k ) locally uniformly in C 1 , 2 ( ( 0 , T ] × [ 0 , ) ) as n ,

At last, combining with the regularity of parabolic equations, we see that (4.7) holds, so the continuity of U ( t , z ; k ) with respect to nonnegative k K satisfying k ¯ [ 0 , c 0 ) is established.

The proof of this theorem is complete now.□

Thanks to the above theorem, we know that for any fixed k K with k ¯ c 0 , problem (4.1) only admits trivial solution 0; while for any nonnegative k K with 0 k ¯ < c 0 , problem (4.1) admits a unique positive solution U ( t , z ; k ) . In order to establish the existence and the uniqueness of time periodic traveling semi-wave, for any t ( τ , T ) and any fixed nonnegative k K , let us define

(4.8) μ [ k ] ( t ) μ U z ( t , 0 ; k ) , 0 k ¯ < c 0 , 0 , k ¯ c 0 ,

where μ > 0 is given in problem (P). Thanks to Theorem 4.1, we can check that the mapping μ is well-defined and show the following result on the existence and uniqueness of solution of k = μ [ k ] .

Lemma 4.2

Under the assumptions of Theorem 4.1, for any fixed μ > 0 , we can find a unique nonnegative function k μ * K with k μ * ¯ ( 0 , c 0 ) such that

k μ * ( t ) = μ [ k * ] ( t ) = μ U z * ( t , 0 ; k μ * ) > 0 f o r t ( τ , T ) ,

where U * ( t , z ; k μ * ) is the unique positive solution of problem (4.1) with k replaced by k μ * .

Proof

Thanks to Theorem 4.1, we see that for any fixed nonnegative function k K satisfying k ¯ [ 0 , c 0 ) , there is a unique positive solution U ( t , z ; k ) to problem (4.1), and it is direct to check that μ [ k ] ( t ) = μ U z ( t , 0 ; k ) is non-increasing in k for t ( τ , T ) .

It is direct to compute that μ [ 0 ] ( t ) = μ U z ( t , 0 ; 0 ) > 0 and μ [ c 0 ] ( t ) = 0 < c 0 for t ( τ , T ) , then we can see that μ [ ] maps S continuously into a precompact set in S , where S is defined by

S { k K : k 0 in [ 0 , T ] and k ¯ c 0 } .

By using the Schauder fixed point theorem we can find some k μ * S such that

k μ * ( t ) = μ [ k μ * ] ( t ) = μ U z * ( t , 0 ; k μ * ) for t ( τ , T ) .

Moreover, one can check that k * ( t ) , 0 in ( τ , T ) and then U z * ( t , 0 ; k * ) , 0 in ( τ , T ) . This, together with Theorem 4.1, infers that

0 k μ * ¯ < c 0 and U z * ( t , 0 ; k μ * ) > 0 for all t ( τ , T ) .

Recalling that k μ * ( t ) = μ [ k μ * ] ( t ) = μ U z * ( t , 0 ; k * ) in ( τ , T ) , thus we obtain that k μ * ( t ) > 0 in ( τ , T ) , which yields the existence of k μ * .

The rest of this proof is devoted to the uniqueness of k μ * , i.e., we shall show that if k i K with k i ¯ [ 0 , c 0 ) and k i ( t ) = μ [ k i ] ( t ) = μ U z ( t , 0 ; k i ) in ( τ , T ) ( i = 1 , 2 ), then k 1 k 2 in ( τ , T ) . Suppose by way of contradiction that k 1 k 2 in ( τ , T ) . First of all, we claim that k 1 , k 2 in ( τ , T ) does not hold. Otherwise, if k 1 , k 2 in ( τ , T ) , it follows from the monotonicity of U ( t , z ; k ) in k that

k 1 ( t ) = μ [ k 1 ] ( t ) = μ U z ( t , 0 ; k 1 ) > μ U z ( t , 0 ; k 2 ) = μ [ k 2 ] ( t ) = k 2 ( t ) in ( τ , T ) ,

which is a contradiction. Similarly, k 2 , k 1 in ( τ , T ) does not hold. Consequently, there are s 1 , s 2 ( τ , T ) such that

(4.9) k 2 ( s 1 ) < k 1 ( s 1 ) , k 2 ( s 2 ) > k 1 ( s 2 ) .

Without loss of generality, we may assume that s 1 < s 2 . Combining this with (4.9), we see

s * sup { s [ s 1 , s 2 ) : k 1 ( t ) > k 2 ( t ) for t [ s 1 , s ) }

is well-defined and s * < s 2 . We then see that k 1 ( t ) > k 2 ( t ) in [ s 1 , s * ) and k 2 ( s * ) = k 1 ( s * ) . Let us define

K 1 ( t ) s 1 t k 1 ( s ) d s , K 2 ( t ) s 1 t k 2 ( s ) d s + X with X s 1 s * [ k 1 ( t ) k 2 ( t ) ] d t .

It is direct to check that K 2 ( t ) > K 1 ( t ) for t [ s 1 , s * ) and K 2 ( s * ) = K 1 ( s * ) (denoted it by x * ). For i = 1 or 2, problem (4.1) with k = k i has a unique positive solution U ( t , z ; k i ) . Let us consider the functions U k i ( t , x ) U ( t , K i ( t ) x ; k i ) for i = 1 , 2 . It is easy to check that

(4.10) U t k i = d U x x k i + U k i ( a b U k i ) , t [ s 1 , s * ] , x < K i ( t ) , U k i ( t , K i ( t ) ) = 0 , U k i ( t , ) = P * ( t ) , t [ s 1 , s * ] , k i ( t ) = μ [ k i ] ( t ) = μ U x k i ( t , K i ( t ) ) , t [ s 1 , s * ] .

Set W ( t , x ) U k 2 ( t , x ) U k 1 ( t , x ) for ( t , x ) Ω [ s 1 , s * ] × ( , K 1 ( t ) ) . Since there is neither seasonal succession nor impulsive harvesting in problem (4.10), it follows from the same argument as in [6, Theorem 2.5] (see also [30, Lemma 3.6]) that W ( t , x ) > 0 in Ω \ { ( s * , x * ) } and W x ( s * , x * ) < 0 . This, together with k i ( t ) = μ U x k i ( t , K i ( t ) ) , yields that k 1 ( s * ) < k 2 ( s * ) , which is in contradiction to the definition of s * , thus k 1 k 2 in ( τ , T ) , as we wanted. The proof of this lemma is complete.□

Next let us consider the dependence of k μ * on the parameter μ and show the following result.

Proposition 4.3

Let c 0 and k μ * be given in Theorem 4.1 and Lemma 4.2, respectively, then we obtain that

  1. k μ * is continuous in μ ;

  2. when 0 < μ 1 < μ 2 ,

    (4.11) k μ 1 * ( t ) k μ 2 * ( t ) i n [ 0 , T ] , a n d k μ 1 * ¯ < k μ 2 * ¯ ;

  3. lim μ k μ * ¯ = c 0 .

Proof

(i) For any given μ > 0 , let μ n μ as n . Recall that

k μ n * ( t ) = μ n U z * ( t , 0 ; k μ n * ) in ( τ , T ) , and k μ n * ( t ) 0 in [ 0 , τ ] .

Thanks to Step 4 in the proof of Theorem 4.1, we see that

(4.12) U * ( t , z ; k μ n * ) U ( t , z ; 0 ) , ( t , z ) [ 0 , T ] × [ 0 , ) , k μ n * ( t ) = μ n U z * ( t , 0 ; k μ n * ) μ n U z ( t , 0 ; 0 ) , t ( τ , T ) ,

where U ( t , z ; 0 ) is the unique positive solution of problem (4.1) with k = 0 . Applying the L p theory to U * ( t , z ; k μ n * ) , we can find p > 1 large such that for any compact subset D of [ 0 , T ] × [ 0 , ) , { U * ( t , z ; k μ n * ) } is bounded in W p 1 , 2 ( D ) . Combining this with the Schauder estimate, we can find a subsequence of { U * ( t , z ; k μ n * ) } (denoted it still by U * ( t , z ; k μ n * ) ) and a function W ( t , z ) C ( 1 + ν ) 2,1 + ν ( ( 0 , T ] × [ 0 , ) ) such that

U * ( t , z ; k μ n * ) W ( t , z ) locally uniformly in C ( 1 + ν ) 2,1 + ν ( ( 0 , T ] × [ 0 , ) ) .

On the one hand, up to a subsequence we see that k μ n * k * weakly in L 2 ( ( τ , T ) ) for some k * with k * 0 in [ 0 , τ ] . Thus, it is direct to check that W is a weak solution of problem (4.1) with k replaced by k * . On the other hand, noting that k μ n * ( t ) 0 for t [ 0 , τ ] and

k μ n * ( t ) = μ n U z * ( t , 0 ; k μ n * ) μ W z ( t , 0 ) in C ν 2 ( ( τ , T ) ) ,

it thus follows that k * ( t ) = μ W z ( t , 0 ) for t ( τ , T ) and k μ n * ( t ) k * ( t ) in C ν 2 ( ( τ , T ) ) .

In what follows let us show that k * ( t ) k μ * ( t ) in [ 0 , T ] . Indeed, if W ( t , z ) 0 in [ 0 , T ] × [ 0 , ) , we have k * ( t ) 0 in [ 0 , T ] , so k μ n * 0 in C ν 2 ( ( τ , T ) ) , which implies that we can find a positive constant ε < c 0 small and an integer n > 1 large such that 0 < k μ n * ( t ) ε < c 0 for t ( τ , T ) . Let k ε ( t ) be an element in the function class K with k ε ( t ) ε for t ( τ , T ) . Combining this with Step 4 in the proof of Theorem 4.1 and the comparison principle, we obtain that U * ( t , z ; k μ n * ) U ( t , z ; k ε ) in [ 0 , T ] × [ 0 , ) for all large n , so

k μ n * ( t ) = μ n U z * ( t , 0 ; k μ n * ) μ n U z ( t , 0 ; k ε ) μ U z ( t , 0 ; k ε ) > 0 for t ( τ , T ) ,

which is in contradiction to k * ( t ) 0 in [ 0 , T ] . Hence, we have proved that W ( t , z ) > 0 in [ 0 , T ] × [ 0 , ) and k * ( t ) = μ W z ( t , 0 ) for t ( τ , T ) . This, together with Lemma 4.2, produces that k * ( t ) k μ * ( t ) for t [ 0 , T ] . Thus, we have proved that k μ n * ( t ) k μ * ( t ) in C ν 2 ( ( τ , T ) ) , which establishes the continuity of k μ * with respect to μ > 0 .

(ii) It follows from the definition of μ and the monotonicity of U z ( t , 0 ; k ) in k that when 0 < μ 1 < μ 2 , and k i K with 0 k 1 , k 2 and k 2 ¯ < c 0 , we see that

(4.13) μ 1 [ k 2 ] ( t ) μ 1 [ k 1 ] ( t ) < μ 2 [ k 1 ] ( t ) for t ( τ , T ) .

It is seen from (4.13) that either k μ 1 * ( t ) , k μ 2 * ( t ) in ( τ , T ) or there are s i ( τ , T ) , i = 1 , 2 such that (4.9) holds, where k μ i * ( t ) is the unique fixed point of μ i . The same arguments as in the proof of Lemma 4.2 can be used to infer that k μ 1 * ( t ) , k μ 2 * ( t ) in ( τ , T ) . Recalling that k μ i * K , then we see that k μ 1 * ( t ) , k μ 2 * ( t ) in [ 0 , T ] . This implies that k μ 1 * ¯ < k μ 2 * ¯ , thus the second assertion of the proposition follows.

(iii) Since k μ * ¯ c 0 for any μ > 0 and k μ * ¯ is increasing in μ > 0 , we have that lim μ k μ * ¯ c 0 . We shall prove that lim μ k μ * ¯ c 0 . Once this is proved, then we see that lim μ k μ * ¯ = c 0 .

In what follows, let us prove lim μ k μ * ¯ c 0 . For any positive sequence { μ n } satisfying μ n as n , it follows from the monotonicity of U * ( t , z ; k μ n * ) in z and the fact that U * ( 0 , ; k μ n * ) = P * ( 0 ) > 0 = U * ( 0 , 1 ; k μ n * ) that we can find a unique z n [ 0 , ] such that

U * ( 0 , z n ; k μ n * ) = 1 2 P * ( 0 ) ,

The comparison principle can be used to infer that z n is monotonically increasing in n , which implies the following limit

lim n z n = z * [ 0 , ] .

Let us define

U n ( t , z ) U * ( t , z + z n ; k μ n * ) for t [ 0 , T ] , z [ z n , ) ,

then the L p theory and the embedding theorem can be used to obtain that there exists a subsequence of U n (still denoted them by { U n } such that U n converges to some function U * locally uniformly in C ( 1 + ν ) 2,1 + ν ( [ 0 , T ] × ( z * , ) ) . At the same time, thanks to the above proof, we see that k μ n * k weakly in L 2 ( ( τ , T ) ) for some element k in the function class K . Thus, it is direct to check that U * ( 0,0 ) = 1 2 P * ( 0 ) , U * ( t , z ) 0 in [ 0 , T ] × ( z * , ) , and U * is a weak solution of problem (4.1) with k replaced by k over [ 0 , T ] × ( z * , ) . Recalling that k μ n * ( t ) 0 in [ 0 , τ ] and k μ n * ( t ) = μ n U z * ( t , 0 ; k μ n * ) in ( τ , T ) , thus we have that

U z * ( t , z * ) = 0 for t ( τ , T ) .

We claim that z * = . Suppose for the contradiction that z * < . As U * ( t , z * ) 0 for t [ 0 , T ] , thus the above discussions yield that U * ( t , z ) = 0 in ( τ , T ) × [ z * , ) . However, it follows from the strong maximum principle and the Hopf lemma that U * ( t , z ) > 0 in ( τ , T ) × ( z * , ) , which is a contradiction. Thus, z * = , which implies that U * ( t , z ) satisfies

U t * = δ U * , t ( 0 + , τ ] , z R , U t * d U z z * + k ( t ) U z * = U * ( a b U * ) , t ( τ , T ] , z R , U z * ( t , z ) > 0 , t [ 0 , T ] , z R , U * ( 0 + , z ) = H ( U * ( 0 , z ) ) , z R , U * ( t , ) = 0 , U * ( t , ) = P * ( t ) , t [ 0 , T ] , U * ( 0,0 ) = 1 2 P * ( 0 ) , U * ( 0 , z ) = U * ( T , z ) , z R .

It is direct to check that ψ ( t , x ) U * t , 0 t k ( s ) d s x is a T -periodic traveling wave solution of the following periodic problem:

(4.14) u t = δ u , t ( 0 + , τ ] , x R , u t = d u x x + a u b u 2 , t ( τ , T ] , x R , u ( 0 + , x ) = H ( u ( 0 , x ) ) , x R , u ( 0 , x ) = u ( T , x ) , x R .

It follows from [16,19,37] that c 0 is the minimum wave speed for the monotone T -periodic traveling wave solutions of problem (4.14), thus we have k ¯ c 0 , which implies that lim μ k μ * ¯ c 0 , as we wanted.

The proof of this proposition is complete.□

4.2 Asymptotic spreading speed

This section deals with the estimate for the asymptotic spreading speed of two free boundaries and gives the proof of Theorem 1.2.

Proof of Theorem 1.2

We only prove (1.6) for h ( t ) , since the similar argument can be used to prove the assertion for g ( t ) .

Let k μ * ( t ) and U * ( t , z ; k μ * ) be given in Lemma 4.2. Noting that spreading happens, then for any ε > 0 , we can find an integer m * > 1 large such that

u ( t , x ) ( 1 + ε ) P * ( t ) for t m * T , x [ g ( t ) , h ( t ) ] .

Recalling that U * ( t , z ; k μ * ) P * ( t ) uniformly for t [ 0 , T ] as z , then we can find L > 0 large such that

( 1 + ε ) U * ( t , z ; k μ * ) > P * ( t ) for t [ 0 , T ] , z L .

Let us define

ξ ( t ) = ( 1 + ε ) 2 0 t k μ * ( s ) d s + L + h ( m * T ) for t 0 , w ( t , x ) = ( 1 + ε ) 2 U * ( t , ξ ( t ) x ; k μ * ) for t 0 , x ξ ( t ) ,

and check that w is a supersolution of problem (P). Put z = z ( t , x ) ξ ( t ) x . We can compute that for x [ g ( m * T ) , h ( m * T ) ] ,

w ( 0 , x ) ( 1 + ε ) 2 U * ( 0 , L ; k μ * ) ( 1 + ε ) P * ( 0 ) = ( 1 + ε ) P * ( m * T ) u ( m * T , x ) ,

and for t ( m T + τ , ( m + 1 ) T ) ,

μ w x ( t , ξ ( t ) ) = μ ( 1 + ε ) 2 U z * ( t , 0 ; k μ * ) = ( 1 + ε ) 2 k μ * ( t ) = ξ ( t ) .

Clearly, w ( t , h ( t ) ) > 0 = w ( t , ξ ( t ) ) for t > 0 . Note that k * ( t ) 0 in ( m T , m T + τ ) , then ξ ( t ) = ξ ( m T ) in ( m T , m T + τ ) . A direct calculation can be used to imply that for t ( ( m T ) + , m T + τ ) and x < ξ ( t ) ,

w t = ( 1 + ε ) 2 [ U t * + ( 1 + ε ) 2 k μ * U z * ] = ( 1 + ε ) 2 U t * = δ w ,

and for t ( m T + τ , ( m + 1 ) T ] and x < ξ ( t ) ,

w t = ( 1 + ε ) 2 [ U t * + ( 1 + ε ) 2 k μ * U z * ] = ( 1 + ε ) 2 [ d U z z * k μ * U z * + U * ( a b U * ) + ( 1 + ε ) 2 k μ * U z * ] ( 1 + ε ) 2 [ d U z z * + U * ( a b U * ) ] d w x x + a w b w 2 ,

where U z * ( t , z ; k μ * ) 0 in ( 0 , ) × [ 0 , ) is used. Moreover, we can check that for x ξ ( t ) ,

w ( ( m T ) + , x ) = ( 1 + ε ) 2 U * ( ( m T ) + , ξ ( m T ) x ; k μ * ) = ( 1 + ε ) 2 H ( U * ( m T , ξ ( m T ) x ; k μ * ) ) H ( ( 1 + ε ) 2 U * ( m T , ξ ( m T ) x ; k μ * ) ) = H ( w ( m T , x ) ) ,

where we have used the monotonicity of H ( s ) s in s 0 . Thus, Lemma 2.1 can be used to show that

u ( t + m * T , x ) w ( t , x ) , h ( t + m * T ) ξ ( t ) for t 0 , x h ( t + m * T ) .

Combining this with the definition of k μ * ( t ) , we see that

limsup t h ( t ) t limsup t ξ ( t m * T ) t = ( 1 + ε ) 2 1 T τ T k μ * ( s ) d s .

Letting ε 0 , we have

(4.15) limsup t h ( t ) t 1 T τ T k μ * ( t ) d t .

On the other hand, we define

η ( t ) = ( 1 ε ) 2 0 t k μ * ( s ) d s + h 0 for t 0 , v ( t , x ) = ( 1 ε ) 2 U * ( t , η ( t ) x ; k μ * ) for t 0 , x ( g ( 0 ) , η ( t ) ] ,

and check that v is a subsolution of problem (P). For all t 0 , it is easy to observe that v ( t , η ( t ) ) = 0 . Set z = z ( t , x ) η ( t ) x . Recalling that spreading happens, then we can find a integer m * large such that

u ( t , x ) ( 1 ε ) P * ( t ) for t m * T , x [ g ( 0 ) , h ( 0 ) ] .

This, together with the fact that U * ( t , z ; k μ * ) P * ( t ) for all t [ 0 , T ] and z 0 , yields that

u ( m * T , x ) ( 1 ε ) P * ( m * T ) ( 1 ε ) 2 U * ( 0 , η ( 0 ) x ; k μ * ) = v ( 0 , x ) for x [ g ( 0 ) , η ( 0 ) ] ,

and

u ( t + m * T , g ( 0 ) ) ( 1 ε ) P * ( t ) > v ( t , g ( 0 ) ) for t > 0 .

When t ( m T + τ , ( m + 1 ) T ] , we check that

μ v x ( t , η ( t ) ) = μ ( 1 ε ) 2 U z * ( t , 0 ; k μ * ) = ( 1 ε ) 2 k μ * ( t ) = η ( t ) .

Noting that k μ * ( t ) 0 for t ( m T , m T + τ ) , then we have η ( t ) = η ( m T ) for t [ m T , m T + τ ) . It is direct to calculate that for ( t , x ) ( ( m T ) + , m T + τ ) × [ g ( 0 ) , η ( t ) ) ,

v t = ( 1 ε ) 2 [ U t * + ( 1 ε ) 2 k μ * U z * ] = ( 1 ε ) 2 U t * = δ v ,

and for ( t , x ) ( m T + τ , ( m + 1 ) T ) × [ g ( 0 ) , η ( t ) ) ,

v t = ( 1 ε ) 2 [ U t * + ( 1 ε ) 2 k μ * U z * ] = ( 1 ε ) 2 [ d U z z * k μ * U z * + U * ( a b U * ) + ( 1 ε ) 2 k μ * U z * ] ( 1 ε ) 2 [ d U z z * + U * ( a b U * ) ] d v x x + a v b v 2 ,

where the fact that U z * ( t , z ; k μ * ) 0 in ( 0 , ) × [ 0 , ) is used. Moreover, a direct calculation can be used to check that for g ( 0 ) x η ( t ) ,

v ( ( m T ) + , x ) = ( 1 ε ) 2 U * ( ( m T ) + , η ( m T ) x ; k μ * ) = ( 1 ε ) 2 H ( U * ( m T , η ( m T ) x ; k μ * ) ) H ( ( 1 ε ) 2 U * ( m T , η ( m T ) x ; k μ * ) ) = H ( v ( m T , x ) ) ,

where the monotonicity of H ( s ) s in s 0 is used. It thus follows from Lemma 2.2 that

u ( t + m * T , x ) v ( t , x ) , h ( t + m * T ) η ( t ) for t 0 , x [ g ( 0 ) , η ( t ) ] .

Thanks to this and the definition of k μ * ( t ) , we see

liminf t h ( t ) t liminf t η ( t m * T ) t = ( 1 ε ) 2 1 T τ T k μ * ( s ) d s .

Setting ε 0 , we obtain that

(4.16) liminf t h ( t ) t 1 T τ T k μ * ( t ) d t .

This, together with (4.15), yields that

lim t h ( t ) t = 1 T τ T k μ * ( t ) d t .

Hence the desired result follows.□

5 Discussion

In the present article, a free boundary model with seasonal succession and impulsive harvesting is studied, and this model can be used to describe the biological invasion of the species in a new environment, whose growth and survival is affected by seasonal succession and impulsive harvesting. The main contribution of this article is the characterization of the dynamical behavior of solutions, which is stated in Theorem 1.1. The second contribution is that we use the T -periodic traveling semi-waves to depict the asymptotic spreading speed of two free boundaries to the spreading solutions (see Theorem 1.2).

The assumption e δ τ a ( T τ ) < H ( 0 ) < 1 is important, otherwise if 0 < H ( 0 ) e δ τ a ( T τ ) (which means that the harvesting rate is too big, or the duration of the bad season is too long, or the season is too bad), it then follows from [21,25] that u 0 as t , thus this case is trivial.

In the current work, we have assumed that the species live in a one-dimensional homogeneous space. However, the habitat of a biological population, in general, can be rather complicated. For instance, natural river systems are often in a spatial network structure such as dendritic trees. The network topology can greatly influence the species persistence and extinction. Therefore, as in [15,26,27], it would be interesting to consider a more general river habitat (bounded or unbounded) consisting of more than one branch. We plan to study these problems with seasonal succession and impulsive harvesting in future work.

Acknowledgments

The authors would like to thank the anonymous referees for their careful reading and helpful comments, which led to an improvement of our original manuscript.

  1. Funding information: This work was supported by Shandong Provincial Natural Science Foundation of China (No. ZR2023YQ002), and the Support Plan for Outstanding Youth Innovation Team in Shandong Higher Education Institutions (No. 2021KJ037).

  2. Author contributions: Both authors contributed equally and significantly to this manuscript, reviewed all the results and approved the final version of the manuscript.

  3. Conflict of interest: The authors state no conflict of interest.

  4. Data availability statement: No data were used in this study.

References

[1] D. G. Aronson and H. F. Weinberger, Multidimensional nonlinear diffusion arising in population genetics, Adv. Math. 30 (1978), 33–76, https://doi.org/10.1016/0001-8708(78)90130-5. Search in Google Scholar

[2] R. J. H. Beverton and S. J. Holt, On the Dynamics of Exploited Fish Populations, Springer, Netherlands, 1993. 10.1007/978-94-011-2106-4Search in Google Scholar

[3] G. Bunting, Y. Du, and K. Krakowski, Spreading speed revisited: analysis of a free boundary model, Netw. Heterog. Media, 7 (2012), 583–603, 10.3934/nhm.2012.7.583. Search in Google Scholar

[4] D. L. DeAngelis, J. C. Trexler, and D. D. Donalson, Competition dynamics in a seasonally varying wetland, Spatial Ecology 1 (2009), 1–13, https://doi.org/10.1201/9781420059861.ch1. Search in Google Scholar

[5] Y. Du, J. Fang, and N. Sun, A delay induced nonlocal free boundary problem, Math. Ann. 386 (2023), 2061–2106, https://doi.org/10.1007/s00208-022-02451-3. Search in Google Scholar

[6] Y. Du, Z. M. Guo, and R. Peng, A diffusive logistic model with a free boundary in time-periodic environment, J. Funct. Anal. 265 (2013), 2089–2142, https://doi.org/10.1016/j.jfa.2013.07.016. Search in Google Scholar

[7] Y. Du and Z. Lin, Spreading-vanishing dichtomy in the diffusive logistic model with a free boundary, SIAM J. Math. Anal. 42 (2010), 377–405, https://doi.org/10.1137/090771089. Search in Google Scholar

[8] Y. Du and B. Lou, Spreading and vanishing in nonlinear diffusion problems with free boundaries, J. Eur. Math. Soc. 17 (2015), 2673–2724, https://doi.org/10.4171/JEMS/568. Search in Google Scholar

[9] Y. Du, H. Matsuzawa, and M. Zhou, Sharp estimate of the spreading speed determined by nonlinear free boundary problems, SIAM J. Math. Anal. 46 (2014), 375–396, https://doi.org/10.1137/130908063. Search in Google Scholar

[10] P. J. DuBowy, Waterfowl communities and seasonal environments: Temporal variabolity in interspecific competition, Ecology 69 (1988), 1439–1453, https://doi.org/10.2307/1941641. Search in Google Scholar

[11] M. Fazly, M. A. Lewis, and H. Wang, On impulsive reaction-diffusion models in higher dimensions, SIAM J. Appl. Math. 77 (2017), 224–246, https://doi.org/10.1137/15M104666. Search in Google Scholar

[12] M. Fazly, M. A. Lewis, and H. Wang, Analysis of propagation for impulsive reaction-diffusion models, SIAM J. Appl. Math. 80 (2020), 521–542, https://doi.org/10.1137/19M124648. Search in Google Scholar

[13] S.-B. Hsu and X.-Q. Zhao, A Lotka-Volterra competition model with seasonal succession, J. Math. Biol. 64 (2012), 109–130, https://doi.org/10.1007/s00285-011-0408-6. Search in Google Scholar PubMed

[14] S. S. Hu and A. J. Tessier, Seasonal succession and the strength of intra- and interspecific competition in a Daphnia assemblage, Ecology 76 (1995), 2278–2294, https://doi.org/10.2307/1941702. Search in Google Scholar

[15] Y. Jin, R. Peng, and J.-P. Shi, Population dynamics in river networks, J. Nonlinear Sci. 29 (2019), 2501–2545, https://doi.org/10.1007/s00332-019-09551-6. Search in Google Scholar

[16] M. A. Lewis and B.-T. Li, Spreading speed, traveling waves, and minimal domain size in impulsive reaction-diffusion models, Bull. Math. Biol. 74 (2012), 2383–2402, https://doi.org/10.1007/s11538-012-9757-6. Search in Google Scholar PubMed

[17] F. Li, X. Liang, and W. Shen, Diffusive KPP equations with free boundaries in time almost periodic environments: II. Spreading speeds and semi-wave solutions, J. Differ. Equ. 261 (2016), 2403–2445, https://doi.org/10.1016/j.jde.2016.04.035. Search in Google Scholar

[18] Y. Li and N. Sun, A free boundary problem with advection and seasonal succession, J. Differ. Equ. 423 (2025), 512–549, https://doi.org/10.1016/j.jde.2025.01.028. Search in Google Scholar

[19] X. Liang and X.-Q. Zhao, Asymptotic speeds of spread and travelingwaves for monotone semiflows with applications, Comm. Pure Appl. Math. 60 (2007), 1–40, https://doi.org/10.1002/cpa.20154. Search in Google Scholar

[20] Z. Lin, A free boundary problem for a predator-prey model, Nonlinearity 20 (2007), 1883–1892, 10.1088/0951-7715/20/8/004. Search in Google Scholar

[21] Y. Meng, J. Ge, and Z. Lin, Dynamics of a free boundary problem modelling species invasion with impulsive harvesting, Discrete Contin. Dyn. Syst. Ser. B 27 (2022), 7689–7720, 10.3934/dcdsb.2022060. Search in Google Scholar

[22] Y. Meng, Z. Lin, and M. Pedersen, Effects of impulsive harvesting and an evolving domain in a diffusive logistic model, Nonlinearity 34 (2021), 7005–7029, 10.1088/1361-6544/ac1f78. Search in Google Scholar

[23] G. Nadin, The principal eigenvalue of a space-time periodic parabolic operator, Ann. Mat. Pura Appl. 188 (2009), 269–295, https://doi.org/10.1007/s10231-008-0075-4. Search in Google Scholar

[24] R. Peng and D. Wei, The periodic-parabolic logistic equation on RN, Discrete Contin. Dyn. Syst. 32 (2012), 619–641, 10.3934/dcds.2012.32.619. Search in Google Scholar

[25] R. Peng and X.-Q. Zhao, The diffusive logistic model with a free boundary and seasonal succession, Discrete Contin. Dyn. Syst. 33 (2013), 2007–2031, 10.3934/dcds.2013.33.2007. Search in Google Scholar

[26] J. M. Ramirez, Population persistence under advection-diffusion in river networks, J. Math. Biol. 65 (2012), 919–942, https://doi.org/10.1007/s00285-011-0485-6. Search in Google Scholar PubMed

[27] J. Sarhad, S. Manifold, and K. E. Anderson, Geometric indicators of population persistence in branching continuous-space networks, J. Math. Biol. 74 (2017), 981–1009, https://doi.org/10.1007/s00285-016-1045-x. Search in Google Scholar PubMed

[28] C. Tian and S. Ruan, A free boundary problem for Aedes aegypti mosquito invasion, Appl. Math. Model. 46 (2017), 203–217, https://doi.org/10.1016/j.apm.2017.01.050. Search in Google Scholar

[29] N. Sun and J. Fang, Propagation dynamics of Fisher-KPP equation with time delay and free boundaries, Calc. Var. Partial Differ. Equ. 58 (2019), Paper No. 148, 38 pp, https://doi.org/10.1007/s00526-019-1599-8. Search in Google Scholar

[30] N. Sun, B. Lou and M. Zhou, Fisher-KPP equation with free boundaries and time-periodic advections, Calc. Var. Partial Differ. Equ. 56 (2017), Paper No. 61, 36 pp, https://doi.org/10.1007/s00526-017-1165-1. Search in Google Scholar

[31] N. Sun and D. Zhang, Dynamical behavior of solutions of a reaction-diffusion-advection model with a free boundary, Z. Angew. Math. Phys., 75 (2024), Paper No. 40, 30 pp, https://doi.org/10.1007/s00033-023-02183-7. Search in Google Scholar

[32] M. X. Wang, A diffusive logistic equation with a free boundary and sign-changing coefficient in time-periodic environment, J. Funct. Anal. 270 (2016), 483–508, https://doi.org/10.1016/j.jfa.2015.10.014. Search in Google Scholar

[33] M. X. Wang, Existence and uniqueness of solutions of free boundary problems in heterogeneous environments, Discrete Contin. Dyn. Syst. Ser. B 24 (2019), 415–421, https://doi.org/10.3934/dcdsb.2018179. Search in Google Scholar

[34] M. X. Wang, Q. Y. Zhang, and X.-Q. Zhao, Dynamics for a diffusive competition model with seasonal succession and different free boundaries, J. Differ. Equ. 285 (2021), 536–582, https://doi.org/10.1016/j.jde.2021.03.006. Search in Google Scholar

[35] M. X. Wang, Q. Y. Zhang, and X.-Q. Zhao, Traveling waves and spreading properties for a reaction-diffusion competition model with seasonal succession, Nonlinearity 35 (2022), 134–169, 10.1088/1361-6544/ac37f6. Search in Google Scholar

[36] Z. Wang, Q. An, and H. Wang, Properties of traveling waves in an impulsive reaction-diffusion model with overcompensation, Z. Angew. Math. Phys. 74 (2023), Paper No. 114, 17 pp, https://doi.org/10.1007/s00033-023-02004-x. Search in Google Scholar

[37] H. F. Weinberger, Long-time behavior of a class of biological models, SIAM J. Math. Anal. 13 (1982), 353–396, https://doi.org/10.1137/0513028. Search in Google Scholar

[38] R. Wu and X.-Q. Zhao, Spatial invasion of a birth pulse population with nonlocal dispersal, SIAM J. Appl. Math. 79 (2019), 1075–1097, https://doi.org/10.1137/18M1209805. Search in Google Scholar

[39] H. Xu, Z. Lin, and H. Zhu, On an age-structured juvenile-adult model with harvesting pulse in moving and heterogeneous environment, J. Differ. Equ. 405 (2024), 36–75, https://doi.org/10.1016/j.jde.2024.05.045. Search in Google Scholar

[40] Q. Y. Zhang and M. X. Wang, A nonlocal diffusion competition model with seasonal succession and free boundaries, Commun. Nonlinear Sci. Numer. Simul. 122 (2023), Art. 107263, https://doi.org/10.1016/j.cnsns.2023.107263. Search in Google Scholar

[41] Q. Y. Zhang and M. X. Wang, The nonlocal dispersal equation with seasonal succession, Proc. Amer. Math. Soc. 152 (2024), 1083–1097, DOI: https://doi.org/10.1090/proc/16626.10.1090/proc/16626Search in Google Scholar

Received: 2025-01-21
Revised: 2025-05-29
Accepted: 2025-06-24
Published Online: 2025-08-13

© 2025 the author(s), published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

  1. Research Articles
  2. Incompressible limit for the compressible viscoelastic fluids in critical space
  3. Concentrating solutions for double critical fractional Schrödinger-Poisson system with p-Laplacian in ℝ3
  4. Intervals of bifurcation points for semilinear elliptic problems
  5. On multiplicity of solutions to nonlinear Dirac equation with local super-quadratic growth
  6. Entire radial bounded solutions for Leray-Lions equations of (p, q)-type
  7. Combinatorial pth Calabi flows for total geodesic curvatures in spherical background geometry
  8. Sharp upper bounds for the capacity in the hyperbolic and Euclidean spaces
  9. Positive solutions for asymptotically linear Schrödinger equation on hyperbolic space
  10. Existence and non-degeneracy of the normalized spike solutions to the fractional Schrödinger equations
  11. Existence results for non-coercive problems
  12. Ground states for Schrödinger-Poisson system with zero mass and the Coulomb critical exponent
  13. Geometric characterization of generalized Hajłasz-Sobolev embedding domains
  14. Subharmonic solutions of first-order Hamiltonian systems
  15. Sharp asymptotic expansions of entire large solutions to a class of k-Hessian equations with weights
  16. Stability of pyramidal traveling fronts in time-periodic reaction–diffusion equations with degenerate monostable and ignition nonlinearities
  17. Ground-state solutions for fractional Kirchhoff-Choquard equations with critical growth
  18. Existence results of variable exponent double-phase multivalued elliptic inequalities with logarithmic perturbation and convections
  19. Homoclinic solutions in periodic partial difference equations
  20. Classical solution for compressible Navier-Stokes-Korteweg equations with zero sound speed
  21. Properties of minimizers for L2-subcritical Kirchhoff energy functionals
  22. Asymptotic behavior of global mild solutions to the Keller-Segel-Navier-Stokes system in Lorentz spaces
  23. Blow-up solutions to the Hartree-Fock type Schrödinger equation with the critical rotational speed
  24. Qualitative properties of solutions for dual fractional parabolic equations involving nonlocal Monge-Ampère operator
  25. Regularity for double-phase functionals with nearly linear growth and two modulating coefficients
  26. Uniform boundedness and compactness for the commutator of an extension of Riesz transform on stratified Lie groups
  27. Normalized solutions to nonlinear Schrödinger equations with mixed fractional Laplacians
  28. Existence of positive radial solutions of general quasilinear elliptic systems
  29. Low Mach number and non-resistive limit of magnetohydrodynamic equations with large temperature variations in general bounded domains
  30. Sharp viscous shock waves for relaxation model with degeneracy
  31. Bifurcation and multiplicity results for critical problems involving the p-Grushin operator
  32. Asymptotic behavior of solutions of a free boundary model with seasonal succession and impulsive harvesting
  33. Blowing-up solutions concentrated along minimal submanifolds for some supercritical Hamiltonian systems on Riemannian manifolds
  34. Stability of rarefaction wave for relaxed compressible Navier-Stokes equations with density-dependent viscosity
  35. Singularity for the macroscopic production model with Chaplygin gas
  36. Global strong solution of compressible flow with spherically symmetric data and density-dependent viscosities
  37. Global dynamics of population-toxicant models with nonlocal dispersals
  38. α-Mean curvature flow of non-compact complete convex hypersurfaces and the evolution of level sets
  39. High-energy solutions for coupled Schrödinger systems with critical growth and lack of compactness
  40. On the structure and lifespan of smooth solutions for the two-dimensional hyperbolic geometric flow equation
  41. Well-posedness for physical vacuum free boundary problem of compressible Euler equations with time-dependent damping
  42. On the existence of solutions of infinite systems of Volterra-Hammerstein-Stieltjes integral equations
  43. Remark on the analyticity of the fractional Fokker-Planck equation
  44. Continuous dependence on initial data for damped fourth-order wave equation with strain term
  45. Unilateral problems for quasilinear operators with fractional Riesz gradients
  46. Boundedness of solutions to quasilinear elliptic systems
  47. Existence of positive solutions for critical p-Laplacian equation with critical Neumann boundary condition
  48. Non-local diffusion and pulse intervention in a faecal-oral model with moving infected fronts
  49. Nonsmooth analysis of doubly nonlinear second-order evolution equations with nonconvex energy functionals
  50. Qualitative properties of solutions to the viscoelastic beam equation with damping and logarithmic nonlinear source terms
  51. Shape of extremal functions for weighted Sobolev-type inequalities
  52. One-dimensional boundary blow up problem with a nonlocal term
  53. Doubling measure and regularity to K-quasiminimizers of double-phase energy
  54. General solutions of weakly delayed discrete systems in 3D
  55. Global well-posedness of a nonlinear Boussinesq-fluid-structure interaction system with large initial data
  56. Optimal large time behavior of the 3D rate type viscoelastic fluids
  57. Local well-posedness for the two-component Benjamin-Ono equation
  58. Self-similar blow-up solutions of the four-dimensional Schrödinger-Wave system
  59. Existence and stability of traveling waves in a Keller-Segel system with nonlinear stimulation
  60. Existence and multiplicity of solutions for a class of superlinear p-Laplacian equations
  61. On the global large regular solutions of the 1D degenerate compressible Navier-Stokes equations
  62. Normal forms of piecewise-smooth monodromic systems
  63. Fractional Dirichlet problems with singular and non-locally convective reaction
  64. Sharp forced waves of degenerate diffusion equations in shifting environments
  65. Global boundedness and stability of a quasilinear two-species chemotaxis-competition model with nonlinear sensitivities
  66. Non-existence, radial symmetry, monotonicity, and Liouville theorem of master equations with fractional p-Laplacian
  67. Global existence, asymptotic behavior, and finite time blow up of solutions for a class of generalized thermoelastic system with p-Laplacian
  68. Formation of singularities for a linearly degenerate hyperbolic system arising in magnetohydrodynamics
  69. Linear stability and bifurcation analysis for a free boundary problem arising in a double-layered tumor model
  70. Hopf's lemma, asymptotic radial symmetry, and monotonicity of solutions to the logarithmic Laplacian parabolic system
  71. Generalized quasi-linear fractional Wentzell problems
  72. Existence, symmetry, and regularity of ground states of a nonlinear Choquard equation in the hyperbolic space
  73. Normalized solutions for NLS equations with general nonlinearity on compact metric graphs
  74. Boundedness and global stability in a predator-prey chemotaxis system with indirect pursuit-evasion interaction and nonlocal kinetics
  75. Review Article
  76. Existence and stability of contact discontinuities to piston problems
Downloaded on 29.1.2026 from https://www.degruyterbrill.com/document/doi/10.1515/anona-2025-0094/html
Scroll to top button