 \documentclass[12pt]{article}
\usepackage[letterpaper, margin=1in]{geometry} % Adjust margins as needed
\usepackage{amsmath, amssymb, amsthm} % Math packages
\usepackage{enumitem} % Customize itemize/enumerate
\usepackage{fancyhdr} % Header and footer customization
\usepackage{lastpage} % To reference the last page number
\usepackage{algorithm, algpseudocode} %for writing code in your homework file
\usepackage{permute}
\usepackage{faktor} % quotient group notation
\usepackage{upgreek} % variants for greek letters
\usepackage{indentfirst}
\usepackage{xltabular}
\usepackage{graphicx}

%if you want to paste sage or python code directly into latex:
\usepackage{listings}
\lstset{
    language=Python,
    basicstyle=\ttfamily,
    keywordstyle=\bfseries,
    commentstyle=\itshape,
    showstringspaces=false,
    frame=single,
    captionpos=b,
    numbers=left,
    numberstyle=\tiny,
    numbersep=5pt
}


% Header and footer setup
\pagestyle{fancy}
\fancyhf{} % Clear default headers and footers
\rhead{Tenacity Murdie, Dana Hawley, Ashely Dayer, Richard Hall} % Your name
\lhead{Supplementary Information} 
\rfoot{Page \thepage\ of \pageref{LastPage}} % Page numbering
\setlength{\headheight}{14.5pt}


% Define theorem-like environments
\newtheorem{theorem}{Theorem}[section]
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{corollary}[theorem]{Corollary}
\newtheorem{definition}[theorem]{Definition}

\DeclareMathOperator{\Aut}{Aut}
\DeclareMathOperator{\Char}{char}
\DeclareMathOperator{\Gal}{Gal}
\DeclareMathOperator{\id}{id}
\DeclareMathOperator{\lcm}{lcm}
\newcommand{\ord}[1]{\lvert #1 \rvert}
\newcommand{\N}{\mathbb{N}}
\newcommand{\Z}{\mathbb{Z}}
\newcommand{\Q}{\mathbb{Q}}
\newcommand{\R}{\mathbb{R}}
\newcommand{\C}{\mathbb{C}}
\let\epsilon\varepsilon

% Title setup
\title{Theoretical Ecology: Supplementary Information to Modeling the Interactive Effects of Predation and Infection in Food-Subsidized Wildlife Populations}
\author{Tenacity Murdie$^{*,1,2}$, Dana Hawley$^{3}$, Ashley Dayer$^{4}$, Richard Hall$^{1,5,6}$}
\date{}
\begin{document}
\setlength\parindent{15pt}
\maketitle

\begin{align*}
&*\text{Corresponding author: tenacitymurdie2030@u.northwestern.edu}\\
&1.\text{ Odum School of Ecology, University of Georgia, 140 E Green St, Athens, GA 30602, USA}\\
&2.\text{ Engineering Sciences and Applied Mathematics, Northwestern University, 2145 Sheridan Rd,}\\
&\text{Evanston, IL 60208, USA}\\
&3.\text{ Department of Biological Sciences, Virginia Tech, Derring Hall, Blacksburg, VA 24060, USA}\\
&4.\text{ Department of Fish and Wildlife Conservation, Virginia Tech, 310 West Campus Dr,}\\
&\text{Blacksburg, VA 24061, USA}\\
&5.\text{ Department of Infectious Diseases, College of Veterinary Medicine, University of Georgia,}\\
&\text{501 D. W. Brooks Drive, Athens, GA 30602, USA}\\
&6.\text{ Center for the Ecology of Infectious Diseases, University of Georgia, 203 D. W. Brooks Drive,}\\
&\text{Athens, GA 30602, USA}
\end{align*}

\section*{S1. Parameterization}

We applied our model to a well-studied host-pathogen system in North America, conjunctivitis caused by the bacterial pathogen \textit{Mycoplasma gallisepticum} in House Finches (\textit{Haemorhous mexicanus}).\ We utilized published field and laboratory experiments, and continent-wide citizen science studies, to parameterize our model (described below for each parameter). \\

\noindent\textbf{Immigration ($\lambda$)}\\

Utilizing the equilibrium equation $0 = \lambda – mN$, we set $N = 16$ based on literature value ranges of wintering finches in the eastern United States (Hosseini et al. 2004; Sydenstricker et al. 2006; Moyers et al. 2018; Birds of the World 2022).\ With $m = 0.03$ $\text{day}^{-1}$ (see below), $\lambda = 0.48$ birds per day.\\

\noindent\textbf{Emigration ($m$)}\\

Jennelle (2007) estimated 0.15-0.3 as the range for weekly probabilities of House Finch emigration from a local study area. Thus, we calculated the daily probability of emigration from a single backyard using the equation $1-(1-x)^\frac{1}{7}$, where x was 0.15 and 0.3. This gave a range of 0.0229-0.0497 $\text{day}^{-1}$. Another study (Davis 2008) said that most birds captured in the fall were seen for up to 3 months, with birds staying in the yard for an average of 79.5 days. Using the equation $\frac{1}{\text{days}}$, we found the range of emigration values to be 0.0111-0.0126 $\text{day}^{-1}$. Therefore, the overall range is 0.0111-0.0497 $\text{day}^{-1}$. Solving for the midpoint value and rounding, the emigration rate is $m = 0.03$ $\text{day}^{-1}$. \\

\noindent\textbf{Natural Mortality ($\mu$)}\\

Birds of the World (2022) gave an overwinter survival range of 39$\%$-63$\%$, while Hosseini et al.\ (2004) and Hurtado (2008) both used a 50$\%$ survival chance over a year. From these values, we used the following equation: $e^{-(\mu*\text{days})} = \text{survival probability}$. We obtained 0.00190-0.00623 as the probability of natural mortality per day. Solving for the midpoint value and rounding, $\mu = 0.004$ $\text{day}^{-1}$. \\

\noindent\textbf{Predation Rate ($\pi$)}\\

Lindström (1989) examined finches in Sweden and found a $\sim$10$\%$ chance of predation per bird over 1.5 months. Using the equation in the paper to calculate an individual finch’s chance of being predated by a hawk per day in a flock of size 10-100, we obtained $\pi = 0.002$ $\text{day}^{-1}$.\\

\noindent\textbf{Increased Predation Risk ($\theta$)}\\

Birds of the World (2022) stated the high wild bird disease mortality was likely due to increased predation, and Kollias et al.\ (2004) described infected birds as lethargic with reduced activity. Additionally, Adelman et al.\ (2017) reported reduced anti-predator behaviors in infected House Finches. This information suggested $\theta > 1$. However, without more information, we decided to use $\theta=1$ (no increased predation) and $\theta=3$ (high increased predation).\\

\noindent\textbf{Disease Mortality ($v$)}\\

Fleming-Davies et al.\ (2018) lists the maximum disease mortality rate as $0.25$ $\text{day}^{-1}$, which is our high virulence value. Their minimum disease induced mortality rate (used in numerical sensitivity analysis) was $0.01$ $\text{day}^{-1}$, which is our low virulence value.\\

\noindent\textbf{Transmission (Contact * Susceptibility = $\alpha\delta$)}\\

Williams et al. (2014) reports transmission rates (measured in $\text{bird}^{-1}\text{day}^{-1}$) of 0.003 and 0.005 for the low virulence strain.\ We took the average value to find a transmission rate of 0.004 $\text{bird}^{-1}\text{day}^{-1}$. Likewise, the authors reported transmission rates of 0.014 and 0.034 for the high virulence strain.\ We took the average value to find a transmission rate of 0.024 $\text{bird}^{-1}\text{day}^{-1}$. To split transmission into contact and susceptibility, we set contact rates ($\alpha$) equal to one contact per bird per day.\\

\noindent\textbf{Recovery ($\gamma$)}\\

Williams et al. (2014) reports recovery rates (measured in $\text{day}^{-1}$) of 0.061 and 0.111 for the low virulence strain.\ We took the average value to find $\gamma = 0.086$ $\text{day}^{-1}$. Likewise, the authors report transmission rates of 0.000 and 0.006 for the high virulence strain.\ We took the average value to find $\gamma = 0.003$ $\text{day}^{-1}$.
\section*{S2. Minimum Predation Rate ($\pi_{min}$) Calculation}

\noindent\textbf{S2.1. Model Equations}\\

The following differential equations track the number of Susceptible (S) and Infected (I) hosts through time, as well as the cumulative number of hosts dying (C) from infection, predation, or natural causes. The parameters that can be influenced by food subsidy effects have a subscript "$F$".\\

\noindent $\frac{dS}{dt} = \lambda_{F} - m_{F}S - \mu_{F}S - \pi_{F}S - \alpha_{F}\delta_{F}SI + \gamma_{F}I$ \\

\noindent $\frac{dI}{dt} = \alpha_{F}\delta_{F}SI - \mu_{F}I - \theta\pi_{F}I - v_{F}I - \gamma_{F}I$ \\

\noindent $\frac{dC}{dt} = \mu_{F}S + \pi_{F}S + \mu_{F}I + \theta\pi_{F} I+ v_{F}I$ \\

\noindent\textbf{S2.2. Pathogen Basic Reproduction Numbers}\\

The pathogen basic reproduction number ($R_{0}$) represents the number of new infections caused by one infectious individual introduced to a wholly susceptible population (Anderson 1982). The expression for pathogen outbreak potential can be derived by setting $\frac{dS}{dt} = 0, I = 0$ and solving for $S_{eq}$, substituting $S_{eq}$ into the following inequality ($\frac{dI}{dt} > 0$), and rearranging.\\

In the absence of predators, the expression for pathogen basic reproduction number is:\\

\noindent$R_{0np} = \frac{\lambda_F\alpha_F\delta_F}{(\mu_F+v_F+\gamma_F)(m_F+\mu_F)}$\\

In the presence of predators, the expression for pathogen basic reproduction number is:\\

\noindent$R_{0p} = \frac{\lambda_F\alpha_F\delta_F}{(\mu_F+v_F+\gamma_F+\theta\pi_F)(m_F+\mu_F+\pi_F)}$\\

\noindent\textbf{S2.3. Minimum Predation Rates}\\

An $R_{0}$ value below one indicates the pathogen is unable to spread and be sustained within a population. To understand under what conditions predators can prevent pathogen invasion, we set $R_{0p} = 1$ and solved for $\pi$ (which we call the minimum predation rate or $\pi_{min}$) for each parameter altered by food subsidy. Then, for each parameter altered by food subsidy, we found the simplest equation that explained the qualitative dependence of $\pi_{min}$ on food subsidy effect size ($F$).\ Note that $\pi_{min}$ is never negative in our model. Instead, $\pi_{min} = 0$ when $R_{0np} < 1$ (i.e., food subsidy alone is sufficient to prevent pathogen invasion).\\

\noindent\textbf{Immigration: $\lambda_{F} = F\lambda$}\\

\noindent$R_{0p} = 1 = \frac{F\lambda\alpha\delta}{(\mu+v+\gamma+\theta\pi_{min})(m+\mu+\pi_{min})}$\\

\noindent $\Rightarrow \pi_{min} = -(\frac{\mu+v+\gamma}{2\theta}+\frac{\mu+m}{2}) + \sqrt{(\frac{\mu+v+\gamma}{2\theta}-\frac{\mu+m}{2})^2+\frac{F\lambda\alpha\delta}{\theta}}$\\

\noindent At large food subsidy effect sizes, $\pi_{min}$ behaves like $\sqrt{F}$.\\

\noindent\textbf{Emigration: $m_{F} = \frac{m}{F}$}\\

\noindent$R_{0p} = 1 = \frac{\lambda\alpha\delta}{(\mu+v+\gamma+\theta\pi_{min})(\frac{m}{F}+\mu+\pi_{min})}$\\

\noindent $\Rightarrow \pi_{min} = -(\frac{\mu+v+\gamma}{2\theta}+\frac{\mu+\frac{m}{F}}{2}) + \sqrt{(\frac{\mu+v+\gamma}{2\theta}-\frac{\mu+\frac{m}{F}}{2})^2+\frac{\lambda\alpha\delta}{\theta}}$\\

\noindent As $F \rightarrow \infty$, $\frac{1}{F^2} \rightarrow 0$ faster than $\frac{1}{F} \rightarrow 0$. Additionally, while it is true that $\pi_{min}$ behaves like $-\frac{1}{F} + \sqrt{C_1-\frac{C_2}{F}}$ (where $C_1$ and $C_2$ are constants) at large food subsidy effect sizes, we observe that $\sqrt{C_1-\frac{C_2}{F}} \thickapprox \sqrt{C_1}$ when $F$ is large. Thus, we will say that $\pi_{min}$ behaves like $-\frac{1}{F}$ for simplicity.\\

\noindent\textbf{Natural Mortality: $\mu_{F} = \frac{\mu}{F}$}\\

\noindent$R_{0p} = 1 = \frac{\lambda\alpha\delta}{(\frac{\mu}{F}+v+\gamma+\theta\pi_{min})(m+\frac{\mu}{F}+\pi_{min})}$\\

\noindent $\Rightarrow \pi_{min} = -(\frac{\frac{\mu}{F}+v+\gamma}{2\theta}+\frac{\frac{\mu}{F}+m}{2}) + \sqrt{(\frac{\frac{\mu}{F}+v+\gamma}{2\theta}-\frac{\frac{\mu}{F}+m}{2})^2+\frac{\lambda\alpha\delta}{\theta}}$\\

\noindent Using a similar argument for $\pi_{min}$ as we did when food subsidy decreased emigration, we will say that $\pi_{min}$ behaves like $-\frac{1}{F}$ for simplicity.\\

\noindent\textbf{Predation Success: $\pi_{F} = F\pi$}\\

\noindent$R_{0p} = 1 = \frac{\lambda\alpha\delta}{(\mu+v+\gamma+F\theta\pi_{min})(m+\mu+F\pi_{min})}$\\

\noindent $\Rightarrow \pi_{min} = -(\frac{\mu+v+\gamma}{2F\theta}+\frac{\mu+m}{2F}) + \sqrt{(\frac{\mu+v+\gamma}{2F\theta}-\frac{\mu+m}{2F})^2+\frac{\lambda\alpha\delta}{F^2\theta}}$\\

\noindent At large food subsidy effect sizes, $\pi_{min}$ behaves like $\frac{1}{F}$.\\

\noindent\textbf{Disease-Induced Mortality: $v_{F} = \frac{v}{F}$}\\

\noindent$R_{0p} = 1 = \frac{\lambda\alpha\delta}{(\mu+\frac{v}{F}+\gamma+\theta\pi_{min})(m+\mu+\pi_{min})}$\\

\noindent $\Rightarrow \pi_{min} = -(\frac{\mu+\frac{v}{F}+\gamma}{2\theta}+\frac{\mu+m}{2}) + \sqrt{(\frac{\mu+\frac{v}{F}+\gamma}{2\theta}-\frac{\mu+m}{2})^2+\frac{\lambda\alpha\delta}{\theta}}$\\

\noindent Using a similar argument for $\pi_{min}$ as we did when food subsidy decreased emigration, we will say that $\pi_{min}$ behaves like $-\frac{1}{F}$ for simplicity.\\

\noindent\textbf{Contact: $\alpha_{F} = F\alpha$}\\

\noindent$R_{0p} = 1 = \frac{F\lambda\alpha\delta}{(\mu+v+\gamma+\theta\pi_{min})(m+\mu+\pi_{min})}$\\

\noindent $\Rightarrow \pi_{min} = -(\frac{\mu+v+\gamma}{2\theta}+\frac{\mu+m}{2}) + \sqrt{(\frac{\mu+v+\gamma}{2\theta}-\frac{\mu+m}{2})^2+\frac{F\lambda\alpha\delta}{\theta}}$\\

\noindent At large food subsidy effect sizes, $\pi_{min}$ behaves like $\sqrt{F}$.\\

\noindent\textbf{Susceptibility: $\delta_{F} =\frac{\delta}{F}$}\\

\noindent$R_{0p} = 1 = \frac{\lambda\alpha\delta}{F(\mu+v+\gamma+\theta\pi_{min})(m+\mu+\pi_{min})}$\\

\noindent $\Rightarrow \pi_{min} = -(\frac{\mu+v+\gamma}{2\theta}+\frac{\mu+m}{2}) + \sqrt{(\frac{\mu+v+\gamma}{2\theta}-\frac{\mu+m}{2})^2+\frac{\lambda\alpha\delta}{F\theta}}$\\

\noindent At large food subsidy effect sizes, $\pi_{min}$ behaves like $\frac{1}{\sqrt{F}}$.\\

\noindent\textbf{Recovery: $\gamma_{F} = F\gamma$}\\

\noindent$R_{0p} = 1 = \frac{\lambda\alpha\delta}{(\mu+v+F\gamma+\theta\pi_{min})(m+\mu+\pi_{min})}$\\

\noindent $\Rightarrow \pi_{min} = -(\frac{\mu+v+F\gamma}{2\theta}+\frac{\mu+m}{2}) + \sqrt{(\frac{\mu+v+F\gamma}{2\theta}-\frac{\mu+m}{2})^2+\frac{\lambda\alpha\delta}{\theta}}$\\

\noindent While it is true that $\pi_{min}$ behaves like $-F + \sqrt{F^2-F}$ at large food subsidy effect sizes, we observe that $\sqrt{F^2-F}=\sqrt{F(F-1)}\thickapprox \sqrt{F^2}$ at large F. Furthermore, we note that $\lim\limits_{F\rightarrow\infty}(-F + \sqrt{F^2})-\lim\limits_{F\rightarrow\infty}(-F + \sqrt{F^2-F})=0-(-\frac{1}{2})=\frac{1}{2}$, which means the behavior at large F differs by a constant. Thus, we will say that $\pi_{min}$ behaves like $-F + \sqrt{F^2}$ (i.e., $\pi_{min}$ has negative slope when F is negative and $\pi_{min}$ is zero when F is nonnegative) for simplicity.
\section*{S3. Daily Death Difference ($DDD$) Calculation}

\noindent\textbf{S3.1. Model Equations}\\

\noindent $\frac{dS}{dt} = \lambda_{F} - m_{F}S - \mu_{F}S - \pi_{F}S - \alpha_{F}\delta_{F}SI + \gamma_{F}I$ \\

\noindent $\frac{dI}{dt} = \alpha_{F}\delta_{F}SI - \mu_{F}I - \theta\pi_{F}I - v_{F}I - \gamma_{F}I$ \\

\noindent $\frac{dC}{dt} = \mu_{F}S + \pi_{F}S + \mu_{F}I + \theta\pi_{F} I+ v_{F}I$ \\

\noindent\textbf{S3.2. Solving for Equilibrium}\\

We set the equations modeling Susceptible (S) and Infected (I) hosts equal to zero (i.e, $\frac{dS}{dt} = 0, \frac{dI}{dt} = 0$) and solved for S and I at equilibrium ($S_{eq}, I_{eq}$) in the presence and absence of disease and predators.\\

\noindent\textbf{Disease, Predators:}\\

\noindent $0 = \lambda_{F} - m_{F}S_{eq} - \mu_{F}S_{eq} - \pi_{F}S_{eq} - \alpha_{F}\delta_{F}S_{eq}I_{eq} + \gamma_{F}I_{eq}$ \\
\noindent $0 = \alpha_{F}\delta_{F}S_{eq}I_{eq} - \mu_{F}I_{eq} - \theta\pi_{F}I_{eq} - v_{F}I_{eq} - \gamma_{F}I_{eq}$ \\

\noindent Thus,\\

\noindent $\alpha_{F}\delta_{F}S_{eq}I_{eq} = \mu_{F}I_{eq} + \theta\pi_{F}I_{eq} + v_{F}I_{eq} + \gamma_{F}I_{eq}$ \\
\noindent$\Rightarrow \alpha_{F}\delta_{F}S_{eq} = \mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F}$ \\
\noindent$\Rightarrow S_{eq} = \frac{\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}}$ \\

\noindent $(\alpha_{F}\delta_{F}S_{eq} - \gamma_{F})I_{eq} = \lambda_{F} - (m_{F} + \mu_{F} + \pi_{F})S_{eq}$ \\
\noindent$\Rightarrow (\alpha_{F}\delta_{F}\frac{\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}} - \gamma_{F})I_{eq} = \lambda_{F} - (m_{F} + \mu_{F} + \pi_{F})\frac{\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}}$ \\
\noindent $\Rightarrow (\mu_{F}+\theta\pi_{F}+v_{F})I_{eq} = \lambda_{F} - (m_{F} + \mu_{F} + \pi_{F})\frac{\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}}$ \\
\noindent $\Rightarrow I_{eq} = \frac{\lambda_{F}}{(\mu_{F}+\theta\pi_{F}+v_{F})} - \frac{(m_{F} + \mu_{F} + \pi_{F})(\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}(\mu_{F}+\theta\pi_{F}+v_{F})}$ \\

\noindent\textbf{Disease, No Predators:}\\

\noindent $\pi_F \equiv 0$ \\

\noindent Thus,\\

\noindent$S_{eq} = \frac{\mu_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}}$ \\
\noindent $I_{eq} = \frac{\lambda_{F}}{(\mu_{F}+v_{F})} - \frac{(m_{F} + \mu_{F})(\mu_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}(\mu_{F}+v_{F})}$ \\

\noindent\textbf{No Disease, Predators:} \\

\noindent $I_{eq} = 0$\\
\noindent $0 = \lambda_{F} - m_{F}S_{eq} - \mu_{F}S_{eq} - \pi_{F}S_{eq}$ \\

\noindent Thus,\\

\noindent $(m_{F} + \mu_{F} + \pi_{F})S_{eq} = \lambda_{F}$ \\
\noindent$\Rightarrow S_{eq} = \frac{\lambda_{F}}{m_{F} + \mu_{F} + \pi_{F}}$ \\

\noindent $\frac{dC}{dt} = \mu_{F}S_{eq} + \pi_{F}S_{eq} + \mu_{F} I_{eq} + \theta\pi_{F} I_{eq}+ v_{F} I_{eq}$ \\

\noindent\textbf{No Disease, No Predators:}\\

\noindent $\pi_F \equiv 0$ \\

\noindent Thus,\\

\noindent$S_{eq} = \frac{\lambda_{F}}{m_{F} + \mu_{F}}$ \\
\noindent $I_{eq} = 0$ \\

\noindent\textbf{S3.3. Host Deaths Per Day}\\

We substituted the equilibrium values of Susceptible ($S_{eq}$) and Infected ($I_{eq}$) hosts (found above) into the equation for daily host deaths ($\frac{dC}{dt}$) to find the number of host deaths per day at equilibrium in the presence and absence of disease and predators.\\

\noindent $\frac{dC}{dt} = \mu_{F}S_{eq} + \pi_{F}S_{eq} + \mu_{F}I_{eq} + \theta\pi_{F} I_{eq}+ v_{F}I_{eq}$ \\

\noindent\textbf{Disease, Predators:}\\

\noindent$S_{eq} = \frac{\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}}$ \\
\noindent $I_{eq} = \frac{\lambda_{F}}{(\mu_{F}+\theta\pi_{F}+v_{F})} - \frac{(m_{F} + \mu_{F} + \pi_{F})(\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}(\mu_{F}+\theta\pi_{F}+v_{F})}$ \\

\noindent Thus,\\

\noindent $\frac{dC}{dt} = (\mu_{F} + \pi_{F})\frac{\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}} + (\mu_{F} + \theta\pi_{F}+ v_{F})(\frac{\lambda_{F}}{(\mu_{F}+\theta\pi_{F}+v_{F})} - \frac{(m_{F} + \mu_{F} + \pi_{F})(\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}(\mu_{F}+\theta\pi_{F}+v_{F})})$ \\
\noindent $\Rightarrow \frac{dC}{dt} = \frac{(\mu_{F} + \pi_{F})(\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}} + \lambda_{F} - \frac{(m_{F} + \mu_{F} + \pi_{F})(\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}}$\\
\noindent $\Rightarrow \frac{dC}{dt} = \lambda_{F} - \frac{m_{F}(\mu_{F} + \theta\pi_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}}$\\

\noindent\textbf{Disease, No Predators:}\\

\noindent $\pi_F \equiv 0$ \\
\noindent$S_{eq} = \frac{\mu_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}}$ \\
\noindent $I_{eq} = \frac{\lambda_{F}}{(\mu_{F}+v_{F})} - \frac{(m_{F} + \mu_{F})(\mu_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}(\mu_{F}+v_{F})}$ \\

\noindent Thus,\\

\noindent $\frac{dC}{dt} = \mu_{F}\frac{\mu_{F} + v_{F} + \gamma_{F}}{\alpha_{F}\delta_{F}} + (\mu_{F} + v_{F})(\frac{\lambda_{F}}{(\mu_{F}+v_{F})} - \frac{(m_{F} + \mu_{F})(\mu_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}(\mu_{F}+v_{F})})$ \\
\noindent $\Rightarrow \frac{dC}{dt} = \frac{\mu_{F}(\mu_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}} + \lambda_{F} - \frac{(m_{F} + \mu_{F})(\mu_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}})$ \\
\noindent $\Rightarrow \frac{dC}{dt} = \lambda_{F} - \frac{m_{F}(\mu_{F} + v_{F} + \gamma_{F})}{\alpha_{F}\delta_{F}}$ \\

\noindent\textbf{No Disease, Predators:}\\

\noindent$S_{eq} = \frac{\lambda_{F}}{m_{F} + \mu_{F} + \pi_{F}}$ \\
\noindent $I_{eq} = 0$ \\

\noindent Thus,\\

\noindent $\frac{dC}{dt} = \frac{\lambda_{F}(\mu_{F} + \pi_{F})}{m_{F} + \mu_{F} + \pi_{F}}$ \\

\noindent\textbf{No Disease, No Predators:}\\

\noindent $\pi_F \equiv 0$ \\
\noindent$S_{eq} = \frac{\lambda_{F}}{m_{F} + \mu_{F}}$ \\
\noindent $I_{eq} = 0$ \\

\noindent Thus,\\

\noindent $\frac{dC}{dt} = \frac{\lambda_{F}\mu_{F}}{m_{F} + \mu_{F}}$\\

\noindent\textbf{S3.4. Daily Death Difference ($DDD$) Calculation}\\

To quantify the effect predators have on cumulative host mortality, we solved for the daily death difference:\\

\noindent DDD = $\frac{dC}{dt}$ at equilibrium with predators - $\frac{dC}{dt}$ at equilibrium without predators\\

Where positive values indicate host deaths per day increase in the presence of predators and negative values indicate host deaths per day decrease in the presence of predators.\\

\noindent\textbf{Immigration: $\lambda_{F} = F\lambda$}\\

\noindent As food subsidy effect size ($F$) increases, both $R_{0p}$ and $R_{0np}$ increase to infinity. 

\noindent\textbf{$\lambda_{F}$: $R_{0p}<R_{0np}<1$}\\

\noindent DDD = $(\frac{F\lambda(\mu + \pi)}{m + \mu + \pi}) - (\frac{F\lambda\mu}{m + \mu})$\\
\noindent $\Rightarrow$ DDD = $\frac{F\lambda\pi m}{(m+\mu+\pi)(m+\mu)}$ \\

\noindent Thus, DDD is always positive and increases linearly (slope = $\frac{\lambda\pi m}{(m+\mu+\pi)(m+\mu)}$) as food subsidy effect size ($F$) increases.\\

\noindent\textbf{$\lambda_{F}$: $R_{0p}<1<R_{0np}$}\\

\noindent DDD = $(\frac{F\lambda(\mu + \pi)}{m + \mu + \pi}) - (F\lambda - \frac{m(\mu+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $\frac{m(\mu+v+\gamma)}{\alpha\delta} - \frac{F\lambda m}{m+\mu+\pi}$\\

\noindent Thus, DDD always decreases linearly (slope = $- \frac{\lambda m}{m+\mu+\pi}$) as food subsidy effect size ($F$) increases. DDD passes through zero when $F = \frac{(\mu+v+\gamma)(m+\mu+\pi)}{\lambda\alpha\delta}$.\\

\noindent\textbf{$\lambda_{F}$: $1<R_{0p}<R_{0np}$}\\

\noindent DDD = $(F\lambda - \frac{m(\mu+\theta\pi+v+\gamma)}{\alpha\delta}) - (F\lambda - \frac{m(\mu+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $-\frac{m\theta\pi}{\alpha\delta}$\\

\noindent Thus, DDD is always negative and does not change with food subsidy effect size ($F$).\\

\noindent\textbf{Emigration: $m_{F} = \frac{m}{F}$}\\

\noindent As food subsidy effect size ($F$) increases, both $R_{0p}$ and $R_{0np}$ increase and saturate at parameter-dependent values that may or may not be greater than 1. \\

$R_{0p} \rightarrow \frac{\alpha\delta\lambda}{(\mu+\pi)(\mu+v+\gamma+\theta\pi)}$ and $R_{0np} \rightarrow \frac{\alpha\delta\lambda}{\mu(\mu+v+\gamma)}$ as $F \rightarrow \infty$\\

\noindent\textbf{$m_{F}$: $R_{0p}<R_{0np}<1$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{\frac{m}{F} + \mu + \pi}) - (\frac{\lambda\mu}{\frac{m}{F} + \mu})$\\
\noindent $\Rightarrow$ DDD = $\frac{\lambda\pi m}{(\frac{m}{F} + \mu + \pi)(m + F\mu)}$\\

\noindent $\frac{DDD}{dF} = \lambda m (\frac{\mu+\pi}{(m+F\mu+F\pi)^2} - \frac{\mu}{(m+F\mu)^2})$\\
\noindent $\Rightarrow$ $\frac{DDD}{dF} = 0$ when $F = \frac{m}{(\mu^2+\mu\pi)^\frac{1}{2}}$\\

\noindent Thus, DDD is always positive and tends to zero as food subsidy effect size ($F$) increases. If $F = \frac{m}{(\mu^2+\mu\pi)^\frac{1}{2}} > 1$, then DDD will reach a maximum at DDD = $\frac{\lambda\pi(\mu^2+\mu\pi)^\frac{1}{2}}{((\mu^2+\mu\pi)^\frac{1}{2}+\mu+\pi)((\mu^2+\mu\pi)^\frac{1}{2}+\mu)}$.\\

\noindent\textbf{$m_{F}$: $R_{0p}<1<R_{0np}$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{\frac{m}{F} + \mu + \pi}) - (\lambda - \frac{\frac{m}{F}(\mu+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $m(\frac{\mu+v+\gamma}{F\alpha\delta}-\frac{\lambda}{m+F\mu+F\pi})$\\

\noindent $\frac{DDD}{dF} = m(\frac{\lambda(\mu+\pi)}{(m+F\mu+F\pi)^2}-\frac{\mu+v+\gamma}{F^2\alpha\delta})$\\
\noindent $\Rightarrow$ $\frac{DDD}{dF} = 0$ when $F = \frac{m((\mu+\pi)(\mu+v+\gamma)+(\alpha\delta\lambda(\mu+\pi)(\mu+v+\gamma))^\frac{1}{2})}{(\mu+\pi)(\alpha\delta\lambda-(\mu+\pi)(\mu+v+\gamma))}$\\

\noindent Thus, DDD tends to zero as food subsidy effect size ($F$) increases. DDD passes through zero when $F = \frac{m(\mu+v+\gamma)}{\alpha\delta\lambda-(\mu+\pi)(\mu+v+\gamma)}$. If $F = \frac{m((\mu+\pi)(\mu+v+\gamma)+(\alpha\delta\lambda(\mu+\pi)(\mu+v+\gamma))^\frac{1}{2})}{(\mu+\pi)(\alpha\delta\lambda-(\mu+\pi)(\mu+v+\gamma))} > 1$, then DDD will reach a minimum. Note that $F$ when DDD reaches a minimum is larger than $F$ when DDD passes through zero.\\

\noindent\textbf{$m_{F}$: $1<R_{0p}<R_{0np}$}\\

\noindent DDD = $(\lambda - \frac{\frac{m}{F}(\mu+\theta\pi+v+\gamma)}{\alpha\delta}) - (\lambda - \frac{\frac{m}{F}(\mu+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $-\frac{m\theta\pi}{F\alpha\delta}$\\

\noindent Thus, DDD is always negative and tends to zero as food subsidy effect size ($F$) increases.\\

\noindent\textbf{Natural Mortality: $\mu_{F} = \frac{\mu}{F}$}\\

\noindent As food subsidy effect size ($F$) increases, both $R_{0p}$ and $R_{0np}$ increase and saturate at parameter-dependent values that may or may not be greater than 1. \\

$R_{0p} \rightarrow \frac{\alpha\delta\lambda}{(m+\pi)(v+\gamma+\theta\pi)}$ and $R_{0np} \rightarrow \frac{\alpha\delta\lambda}{m(v+\gamma)}$ as $F \rightarrow \infty$\\

\noindent\textbf{$\mu_{F}$: $R_{0p}<R_{0np}<1$}\\

\noindent DDD = $(\frac{\lambda(\frac{\mu}{F} + \pi)}{m + \frac{\mu}{F} + \pi}) - (\frac{\lambda\frac{\mu}{F}}{m + \frac{\mu}{F}})$\\
\noindent $\Rightarrow$ DDD = $\frac{\lambda\pi m}{(m + \frac{\mu}{F} + \pi)(m + \frac{\mu}{F})}$\\

\noindent $\frac{DDD}{dF} = \lambda\mu m (\frac{1}{(Fm+\mu)^2} - \frac{1}{(\mu+Fm+F\pi)^2}) > 0$\\

\noindent Thus, DDD is always positive and increases towards DDD = $\frac{\lambda\pi}{m+\pi}$ as food subsidy effect size ($F$) increases.\\

\noindent\textbf{$\mu_{F}$: $R_{0p}<1<R_{0np}$}\\

\noindent DDD = $(\frac{\lambda(\frac{\mu}{F} + \pi)}{m + \frac{\mu}{F} + \pi}) - (\lambda - \frac{m(\frac{\mu}{F}+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $m(\frac{\mu+Fv+F\gamma}{F\alpha\delta}-\frac{F\lambda}{Fm+\mu+F\pi})$\\

\noindent $\frac{DDD}{dF} = -\mu m (\frac{1}{(F^2\alpha\delta)} + \frac{\lambda}{(\mu+Fm+F\pi)^2}) < 0$\\

\noindent Thus, DDD decreases towards DDD = $\frac{m(v+\gamma)}{\alpha\delta}-\frac{\lambda m}{m+\pi}$ as food subsidy effect size ($F$) increases. DDD passes through zero when $F = \frac{\mu((m+v+\gamma+\pi)+((m+v+\gamma+\pi)^2+4(\alpha\delta\lambda-(m+\pi)(v+\gamma)))^\frac{1}{2})}{2(\alpha\delta\lambda-(m+\pi)(v+\gamma))}$.\\

\noindent\textbf{$\mu_{F}$: $1<R_{0p}<R_{0np}$}\\

\noindent DDD = $(\lambda - \frac{m(\frac{\mu}{F}+\theta\pi+v+\gamma)}{\alpha\delta}) - (\lambda - \frac{m(\frac{\mu}{F}+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $-\frac{m\theta\pi}{\alpha\delta}$\\

\noindent Thus, DDD is always negative and does not change with food subsidy effect size ($F$).\\

\noindent\textbf{Predation Success: $\pi_{F} = F\pi$}\\

\noindent As food subsidy effect size ($F$) increases, $R_{0np}$ remains unchanged while $R_{0p}$ decreases towards zero. \\

\noindent\textbf{$\pi_{F}$: $R_{0p}<R_{0np}<1$}\\

\noindent DDD = $(\frac{\lambda(\mu + F\pi)}{m + \mu + F\pi}) - (\frac{\lambda\mu}{m + \mu})$\\
\noindent $\Rightarrow$ DDD = $\frac{F\lambda\pi m}{(m + \mu + F\pi)(m + \mu)}$\\

\noindent $\frac{DDD}{dF} = \frac{\lambda\pi m}{(m+\mu+F\pi)^2} > 0$\\

\noindent Thus, DDD is always positive and increases towards DDD = $\frac{\lambda m}{m+\mu}$ as food subsidy effect size ($F$) increases.\\

\noindent\textbf{$\pi_{F}$: $R_{0p}<1<R_{0np}$}\\

\noindent DDD = $(\frac{\lambda(\mu + F\pi)}{m + \mu + F\pi}) - (\lambda - \frac{m(\mu+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $m(\frac{\mu+v+\gamma}{\alpha\delta}-\frac{\lambda}{m+\mu+F\pi})$\\

\noindent $\frac{DDD}{dF} = \frac{\lambda\pi m}{(m+\mu+F\pi)^2} > 0$\\

\noindent Thus, DDD increases towards DDD = $\frac{m(\mu+v+\gamma)}{\alpha\delta}$ as food subsidy effect size ($F$) increases. DDD passes through zero when $F = \frac{\lambda\alpha\delta - (m + \mu)(\mu+v+\gamma)}{\pi(\mu+v+\gamma)}$.\\

\noindent\textbf{$\pi_{F}$: $1<R_{0p}<R_{0np}$}\\

\noindent DDD = $(\lambda - \frac{m(\mu+F\theta\pi+v+\gamma)}{\alpha\delta}) - (\lambda - \frac{m(\mu+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $-F\frac{m\theta\pi}{\alpha\delta}$\\

\noindent Thus, DDD is always negative and decreases linearly (slope = $-\frac{m\theta\pi}{\alpha\delta}$) with food subsidy effect size ($F$).\\

\noindent\textbf{Disease-Induced Mortality: $v_{F} = \frac{v}{F}$}\\

\noindent As food subsidy effect size ($F$) increases, both $R_{0p}$ and $R_{0np}$ increase and saturate at parameter-dependent values that may or may not be greater than 1. \\

$R_{0p} \rightarrow \frac{\alpha\delta\lambda}{(m+\mu+\pi)(\mu+\gamma+\theta\pi)}$ and $R_{0np} \rightarrow \frac{\alpha\delta\lambda}{(m+\mu)(\mu+\gamma)}$ as $F \rightarrow \infty$\\

\noindent\textbf{$v_{F}$: $R_{0p}<R_{0np}<1$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{m + \mu + \pi}) - (\frac{\lambda\mu}{m + \mu})$\\
\noindent $\Rightarrow$ DDD = $\frac{\lambda\pi m}{(m + \mu + \pi)(m + \mu)}$\\

\noindent Thus, DDD is always positive and does not change with food subsidy effect size ($F$).\\

\noindent\textbf{$v_{F}$: $R_{0p}<1<R_{0np}$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{m + \mu + \pi}) - (\lambda - \frac{m(\mu+\frac{v}{F}+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $m(\frac{\mu+\frac{v}{F}+\gamma}{\alpha\delta}-\frac{\lambda}{m+\mu+\pi})$\\

\noindent $\frac{DDD}{dF} = -\frac{mv}{F^2\alpha\delta} < 0$\\

\noindent Thus, DDD decreases towards DDD = $m(\frac{\mu+\gamma}{\alpha\delta}-\frac{\lambda}{m+\mu+\pi})$ as food subsidy effect size ($F$) increases. DDD passes through zero when $F = \frac{v(m+\mu+\pi)}{\lambda\alpha\delta - (\mu+\gamma)(m+\mu+\pi)}$.\\

\noindent\textbf{$v_{F}$: $1<R_{0p}<R_{0np}$}\\

\noindent DDD = $(\lambda - \frac{m(\mu+\theta\pi+\frac{v}{F}+\gamma)}{\alpha\delta}) - (\lambda - \frac{m(\mu+\frac{v}{F}+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $-\frac{m\theta\pi}{\alpha\delta}$\\

\noindent Thus, DDD is always negative and does not change with food subsidy effect size ($F$).\\

\noindent\textbf{Contact: $\alpha_{F} = F\alpha$}\\

\noindent As food subsidy effect size ($F$) increases, both $R_{0p}$ and $R_{0np}$ increase to infinity. 

\noindent\textbf{$\alpha_{F}$: $R_{0p}<R_{0np}<1$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{m + \mu + \pi}) - (\frac{\lambda\mu}{m + \mu})$\\
\noindent $\Rightarrow$ DDD = $\frac{\lambda\pi m}{(m+\mu+\pi)(m+\mu)}$ \\

\noindent DDD is always positive and does not change with food subsidy effect size ($F$).\\

\noindent\textbf{$\alpha_{F}$: $R_{0p}<1<R_{0np}$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{m + \mu + \pi}) - (\lambda - \frac{m(\mu+v+\gamma)}{F\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $\frac{m(\mu+v+\gamma)}{F\alpha\delta} - \frac{\lambda m}{m+\mu+\pi}$\\

\noindent $\frac{DDD}{dF} = -\frac{m(\mu+v+\gamma)}{F^2\alpha\delta} < 0$\\

\noindent Thus, DDD decreases towards DDD = $- \frac{\lambda m}{m+\mu+\pi}$ as food subsidy effect size ($F$) increases. DDD passes through zero when $F = \frac{(\mu+v+\gamma)(m+\mu+\pi)}{\lambda\alpha\delta}$.\\

\noindent\textbf{$\alpha_{F}$: $1<R_{0p}<R_{0np}$}\\

\noindent DDD = $(\lambda - \frac{m(\mu+\theta\pi+v+\gamma)}{F\alpha\delta}) - (\lambda - \frac{m(\mu+v+\gamma)}{F\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $-\frac{m\theta\pi}{F\alpha\delta}$\\

\noindent $\frac{DDD}{dF} = \frac{m\theta\pi}{F^2\alpha\delta} > 0$\\

\noindent Thus, DDD is always negative and increases towards zero with food subsidy effect size ($F$).\\

\noindent\textbf{Susceptibility: $\delta_{F} = \frac{\delta}{F}$}\\

\noindent As food subsidy effect size ($F$) increases, both $R_{0p}$ and $R_{0np}$ decrease to zero. 

\noindent\textbf{$\delta_{F}$: $R_{0p}<R_{0np}<1$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{m + \mu + \pi}) - (\frac{\lambda\mu}{m + \mu})$\\
\noindent $\Rightarrow$ DDD = $\frac{\lambda\pi m}{(m+\mu+\pi)(m+\mu)}$ \\

\noindent Thus, DDD is always positive and does not change with food subsidy effect size ($F$).\\

\noindent\textbf{$\delta_{F}$: $R_{0p}<1<R_{0np}$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{m + \mu + \pi}) - (\lambda - F\frac{m(\mu+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $F\frac{m(\mu+v+\gamma)}{\alpha\delta} - \frac{\lambda m}{m+\mu+\pi}$\\

\noindent Thus, DDD increases linearly (slope = $\frac{m(\mu+v+\gamma)}{\alpha\delta}$) as food subsidy effect size ($F$) increases. DDD passes through zero when $F = \frac{\lambda\alpha\delta}{(\mu+v+\gamma)(m+\mu+\pi)}$.\\

\noindent\textbf{$\delta_{F}$: $1<R_{0p}<R_{0np}$}\\

\noindent DDD = $(\lambda - F\frac{m(\mu+\theta\pi+v+\gamma)}{\alpha\delta}) - (\lambda - F\frac{m(\mu+v+\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $-F\frac{m\theta\pi}{\alpha\delta}$\\

\noindent Thus, DDD is always negative and decreases linearly (slope = $-\frac{m\theta\pi}{\alpha\delta}$) with food subsidy effect size ($F$).\\

\noindent\textbf{Recovery: $\gamma_{F} = F\gamma$}\\

\noindent As food subsidy effect size ($F$) increases, both $R_{0p}$ and $R_{0np}$ decrease toward zero. 

\noindent\textbf{$\gamma_{F}$: $R_{0p}<R_{0np}<1$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{m + \mu + \pi}) - (\frac{\lambda\mu}{m + \mu})$\\
\noindent $\Rightarrow$ DDD = $\frac{\lambda\pi m}{(m+\mu+\pi)(m+\mu)}$ \\

\noindent Thus, DDD is always positive and does not change with food subsidy effect size ($F$).\\

\noindent\textbf{$\gamma_{F}$: $R_{0p}<1<R_{0np}$}\\

\noindent DDD = $(\frac{\lambda(\mu + \pi)}{m + \mu + \pi}) - (\lambda - \frac{m(\mu+v+F\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $\frac{m(\mu+v+F\gamma)}{\alpha\delta} - \frac{\lambda m}{m+\mu+\pi}$\\

\noindent Thus, DDD increases linearly (slope = $\frac{m\gamma}{\alpha\delta}$) as food subsidy effect size ($F$) increases. DDD passes through zero when $F = \frac{\lambda\alpha\delta-(\mu+v)(m+\mu+\pi)}{\gamma(m+\mu+\pi)}$.\\

\noindent\textbf{$\gamma_{F}$: $1<R_{0p}<R_{0np}$}\\

\noindent DDD = $(\lambda - \frac{m(\mu+\theta\pi+v+F\gamma)}{\alpha\delta}) - (\lambda - \frac{m(\mu+v+F\gamma)}{\alpha\delta})$\\
\noindent $\Rightarrow$ DDD = $-\frac{m\theta\pi}{\alpha\delta}$\\

\noindent Thus, DDD is always negative and does not change with food subsidy effect size ($F$).\\

\end{document}