Ecological Modelling 275 (2014) 37–47 Contents lists available at ScienceDirect Ecological Modelling jo ur nal ho me page: www.elsevier.com/locate/ecolmodel A Daphnia population model that considers pesticide exposure and demographic stochasticity Richard A. Erickson a,b,∗ , Stephen B. Cox a,c , Jessica L. Oates a,c , Todd A. Anderson a , Christopher J. Salice a , Kevin R. Long d a Department of Environmental Toxicology, The Institute of Environmental and Human Health, Texas Tech University, Lubbock, TX, USA b Upper Midwest Environmental Science Center, United States Geological Survey (CNTS), La Crosse, WI, USA c Research and Testing Laboratory, LLC, Lubbock, TX, USA d Department of Mathematics and Statistics, Texas Tech University, Lubbock, TX, USA a r t i c l e i n f o Article history: Received 9 October 2013 Received in revised form 14 December 2013 Accepted 16 December 2013 Available online 9 January 2014 Keywords: Population modeling Stochastic Demographic stochasticity Ecotoxicology a b s t r a c t Population models have emerged as a powerful tool to better understand the ecological effects of tox- icant exposure. Currently, most ecotoxicology population models are deterministic and fail to account for natural variability in biological processes and uncertainty in parameter estimates. We developed, parameterized, and analyzed a Daphnia population model with three different levels of demographic stochasticity to examine how a pesticide, pendimethalin, affects population dynamics. We conducted laboratory studies to generate the data used for the modeling process. The simplest model only included parameter uncertainty and variability. The second model included daily stochastic fecundities. The third model included stochastic fecundities and stochastic mortalities. Of the three models, the second model with stochastic fecundity best described our laboratory test system. All three models were used to test hypotheses about how pesticides would affect population dynamics. We found that pendimethalin either decreased the baseline juvenile survivorship rate or the carrying capacity. We could differentiate the two test effects with our system. Our findings demonstrate how stochastic population models may provide insight into pesticide exposure. © 2013 Elsevier B.V. All rights reserved. 1. Introduction Ecotoxicology combines principles from ecology and environ- mental toxicology and is the study of chemical toxicants on natural systems (Cairns, 1988; Chapman, 2002). One application of eco- toxicology, ecological risk assessment (ERA), assesses the risk of chemical exposure and other stressors to specific ecological recep- tors and the environment in support of a regulatory framework such as the Federal Insecticide, Fungicide, and Rodenticide Act (Suter, 2008). The most commonly used ERA approaches include quotient-based risk estimates, safety factors, statistical extrap- olation among species, and extrapolation from sub-organismal endpoints (Forbes et al., 2009). These methods provide insight into potential toxicity, but lack ecological relevance and can ignore other important processes for populations (Chapman, 2002). This lack of ecological realism has been acknowledged within the lit- erature and other methods have been proposed to create more ∗ Corresponding author at: Department of Environmental Toxicology, The Insti- tute of Environmental and Human Health, Texas Tech University, Lubbock, TX, USA. Tel.: +1 8068854567. E-mail addresses: raerickson@gmail.com, rerickson@usgs.gov (R.A. Erickson). ecologically relevant outputs and insights (Taub, 1997b; Landis, 2009; Forbes et al., 2011). One such method uses population mod- els to understand and evaluate the effects of toxicant exposure (Pastorok et al., 2002; Barnthouse et al., 2008; Forbes et al., 2011). Population models hold great potential to enhance understand- ing of toxicological effects, but have only seen limited application in ecotoxicology, although the number of models has been increas- ing through time (Calow et al., 1997; Forbes and Forbes, 1994; Forbes et al., 2009, 2010, 2011; Hanson and Stark, 2012). These models are an improvement over traditional approaches because they consider important ecological information and provide insight into population-level effects that are not apparent from focusing on individual and lower levels of biological organization (Newman and Clements, 2008). Furthermore, approaches focused on the popula- tion level provide excellent opportunity for the incorporation of additional ecological processes that are known or thought to be important (Forbes and Forbes, 1994; Hendriks and Van De Guchte, 1997; Taub, 1997a,b; Clements and Newman, 2002; Rohr et al., 2006; Newman and Clements, 2008; Clements and Rohr, 2009). Here, we specifically incorporated three important ecological or methodological processes into our model in an effort to capture salient biological properties of populations: (1) the parameteriza- tion of models using data from a population study with both density 0304-3800/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ecolmodel.2013.12.015