Global Dynamics of a New Huanglongbing Transmission Model with Quarantine Measures

An epidemic model which describes Huanglongbing transmission is pro-posed with the goal of investigating the effect of quarantine measures on the spread of diseases. First of all, the analytical formula for the basic reproduction number 0  is obtained by the means of next generation matrix, and the existence of disease-free equilibrium and endemic equilibrium is discussed. Then, the local stability and the global stability of equilibria are investigated by using Routh-Hurwitz criterion and Lyapunov function, respectively. Numerical simulations indicate that comprehensive quarantine measures can effectively control the spread of Huanglongbing. It provides a reliable tactic ba-sis for preventing the epidemic


Introduction
The citrus Huanglongbing (HLB), allied with the bacterium Candidatus Liberibacter asiaticus (CLas), poses a great threat to the industry of citrus worldwide [1] [2]. The most vital method of disease breadth occurs by the citrus psyllids that serve as vectors [1]. The citrus psyllids consume sap from an infected host through their stylets, the viruses in the sap enter the salivary glands, circulate within the psyllid, and then cause infection between citrus trees and psyllids [3] [4] [5] [6] [7]. The citrus tree infected HLB shows yellow shoots, leaves with blotchy mottle, small lopsided fruits, overall stunting of the trees, in server cases, and ultimately the death of the entire tree [8] [9] [10]. In 2020, citrus Huanglongbing was included in the List of Crop pests and diseases by the Ministry of Agri-culture and Rural Affairs [11].
There are no economically viable curative methods for citrus trees infected HLB due to absence of resistant or tolerant commercial citrus varieties [12]. However, a large number of researchers have studied this disease biologically and come to describe the dynamics of disease mathematically [1] [13]. Recent works have provided different insights into the dynamic mechanism of HLB propagation through differential equation modeling, statistical analysis and individual-based modeling [10]. Jacobsen et al. established a 6-dimensional dynamic model within a citrus orchard and analyzed how the number of each class changes over time due to bacterial transmission between citrus trees and psyllids [7]. Parry et al. fitted a spatially explicit disease model, using specialized statistical methods and software, and then obtained the effect of tree age on transmission parameters, which could be used to predict the spread of early epidemics [14]. Lee et al. [2] conducted a mathematical model that combined experiments and individual-based. It emphasized the need to keep the psyllid populations as low as possible due to the whole grove to be infected before the first symptoms appear on any tree. In fact, removing infected trees and quarantining them are also important measures to prevent as many trees from becoming infected. Specially, if HLB is not yet present, quarantine measures should be enforced to keep it out [15]. The quarantine measures, which began on June 17, 2010, were designed to prevent the transportation of citrus in Florida [7]. Therefore, motivated by these researches, we propose a dynamic model of HLB with quarantine measures for exposed and infected citrus trees, and then perform some theoretical analysis of its properties.
The organization of this paper is as follows: the model is developed and the feasible region, the basic reproduction number and the equilibria of the model are obtained in Section 2. By the Routh-Hurwitz criterion, the local stability of disease-free equilibrium and endemic equilibrium are proved in Section 3. By constructing a suitable Lyapunov function, the global stability of the equilibria is studied in Section 4. In addition, numerical simulations are illustrated in our theoretical results in Section 5 and a brief conclusion is given in the last section.

Model and Preliminaries
The dynamics of the model with a fraction of quarantined citrus trees involve citrus-psyllid interactions, therefore, the citrus tree population given as quarantined host ( ) Q t and removed host ( ) R t , at any time t. Thus, . We assume that the citrus population grows at a recruitment rate h Λ , moreover, 1 q and 2 q represent the rate of the infectious convert to the quarantined host and the rate of exposed to the quarantined host, respectively. According to the principle of the compartmental model, the model with first order nonlinear ordinary differential equations is Applied Mathematics constructed as follows: where the parameters involved in the model (2.1) and their values are showed in Table 1. The schematic diagram for the flow of the dynamics is depicted in Figure 1. In the subsequent section, we will explore the basic qualitative properties of the model (2.1).

Feasible Region Denote
Lemma 1. The feasible region is given by 0 Ω is a positively invariant and globally attractive set for system (2.1).
Proof. Clearly, the solution set    [10] θ Infection intensity relative to infected This implies that lim sup , lim sup .
Thus, all solutions in Ω are uniformly bounded. This completes our proof. = + + + , and consider the following equivalent system of (2.1): It is easy to see is a positively invariant and globally attractive set for system (2.6). From now on, we restrict the analysis of system (2.6) to the positive invariant set 1 Ω . Clearly, the disease free equilibrium (DFE) of system (2.6) exists and is given Now, we calculate the basic reproductive number 0 R of system (2.6) by using the next generation operator approach [19]. Considering the infectious compartments represent the rate of appearance of new infection and the rates of transfer into and out of the class of infected states, respectively. Then the Jacobian matrixes of system (2.6) with respect to 0 P are given below: Using the next generation matrix theory, the expression of the basic reproduction number 0  can be obtained as Remark 1:  indicates the number of secondary infections by the other circular transmission path from I to Y and back to I.

The Existence of Endemic Equilibrium
Next, we consider the existence of endemic equilibrium. Let the right side of equations of system (2.6) be equal to 0, then we obtain algebraic equations as follows: From the third equation of system (2.10), we get By adding the first second equations and the last two equations of system (2.10) respectively, we find Substituting Equations (2.11) and (2.12) into the second and fifth equation of (2.10), we have , for system (2.6) there exists a unique endemic equilibrium * P in the interior of 1 Ω . Therefore, from the above calculation, we get the following theorem.

Local Stability of Equilibria
In this section, we will investigate the local stability of the disease-free equilibrim Proof. The Jacobian matrix of system (2.6) at disease-free equilibrium The characteristic equation of matrix 0  is the following polynomial equa- It is obviously that two eigenvalues of λ µ λ µ λ µ λ µ λ µ µ µ µ λ µ θλ µ θ σ µ θ σ Clearly, there is a negative eigenvalue 1 , the other four eigenvalues satisfy the following quartic h

Global Stability of Equilibria
In this section, we will discuss the global stability of the DFE 0 P and the endemic equilibrium * P .
Theorem 5. The disease-free equilibrium 0 P of system (2.6) is globally asymptotically stable (GAS) in the interior of the set 1 Ω provided that 0 1 ≤  .
Proof. Constructing Lyapunov function then the derivative of (4.1) with respect to t along the solution curves of (2.6) is given by , and hence by the La-Salle's invariance principle [20], we conclude that all trajectories starting in 1 Ω approach 0 P for 0 1 ≤  . That is to say, 0 P is globally asymptotically stable in Next, we will investigate the global stability of the endemic equilibrium * P in the positively invariant set 1 Ω by constructing Lyapunov function.
Theorem 6. The endemic equilibrium * P of system (2.6) is globally asymptotically stable in the interior of the set 1 Proof. From (2.10), we know that * * * * * , , , ,  To illustrate the uniqueness and global stability of * P , we set ( ) . Then the derivative of functions Define Lyapunov function as follows: That is, is the singleton { } * P . Therefore, by the LaSalle's invariance principle [20], it follows that the endemic equilibrium

Numerical Simulation
In this section, we first provide results from numerical simulations of model (2.1) that illustrate and support our theoretical results. In the model (2.1), all parameters are in months and their values are shown in Table 1. According to the above parameters of the model (2.1), we conduct a global sensitivity analysis on the basic reproductive number 0  by employing Latin Hypercube Sampling (LHS) and partial rank correlation coefficients (PRCCs) [22] [23]. Figure 2 and Figure 3 depict our sensitivity and uncertainty analysis, which involved computing the PRCCs of 0  using the LHS method [22]. From Figure 2, we help control the spread of the disease. For uncertainty analysis, it can be seen from Figure 3 that the probability of 0 1 >  is significantly higher than that of , there are still endemic disease ( Figure 5). One interesting thing is that for 0 1 >  , the number of the infected populations has a distinct wavy pattern, with multiple peaks and troughs, which means that at some time the infected individual goes to zero and that doesn't mean that the epidemic is under control, it's going to erupt in the near future. Further, if 1 0.45 q = , 2 0.45 q = , , the endemic disease disappears and the disease is well controlled (Figure 6). This shows that the quarantine measure is effective in controlling the spread of HLB.
In addition, the rate of the exposed to the quarantined ( 1 q ) and rate of the infectious convert to the quarantined ( 2 q ) also have a certain influence on 0  .
Therefore, in order to further analyze the influence of 1 q and 2 q on 0  , the sensitivity of parameters 1 2 , q q for 0  is shown in Figure 7. It can be seen that controlling 2 q alone cannot make 0  drop below 1, and controlling 1 q   alone can only make 0  close to 1, but the joint control of 1 q , 2 q can make According to PRCC analysis (see Figure 2), virus transmission probability from psyllid to plant ( 1 β ), virus transmission probability from plant to psyllid ( 2 β ), natural death rate of citrus population ( h µ ) and natural death rate of psyllid population ( v µ ) have great influence on 0  . We conduct sensitivity analysis of these four parameters to 0  , as shown in Figure 8 and Figure 9. It implies that disease control can be achieved by inhibiting 1 β and 2 β or increasing h µ and v µ simultaneously.

Conclusions
In this work, a citrus-psyllid dynamic model with quarantine measure is formulated. Based on the method of next-generation matrix, we obtain the expression of basic reproductive number 0  . The global stability of disease-free equilibrium and endemic equilibrium are demonstrated by constructing Lyapunov functions. Numerical simulation shows that when 0 1 >  , the number of infected hosts has obvious wavy, with multiple peaks and troughs, which implies that at a time infected individuals tending to zero does not mean that the disease is under control, instead, it may break out soon. Additionally, our investigations show that a certain degree quarantine measure is effective, and it is more effective to isolate the exposed than the infected host. Moreover, comprehensive quarantine can more effectively control the outbreak of disease. These results can provide a reference for fruit industry to conduct comprehensive management of HLB.