Wat. Res. Vol. 31, No. 2, pp. 211-218, 1997 Pergamon Copyright© 1996Publishedby Elsevier ScienceLtd Printed in Great Britain.All rights reserved PII: S0tM3-1354(96)00255-2 0o43-1354/97 $17.00+ 0.oo MODELLING FISH MORTALITY DUE TO URBAN STORM RUN-OFF: INTERACTING EFFECTS OF HYPOXIA AND UN-IONIZED AMMONIA HI~LI~NE MAGAUD, BERNARD MIGEON, PHILIPPE MORFIN, JEANNE GARRIC and ERIC VINDIMIAN Laboratoire d'6cotoxicologie, CEMAGREF, 3 bis quai Chauveau, 69336 Lyon Cedex 09, France (First received May 1994; accepted in revisedform July 1996) Abstract--Storm run-off is a source of acute toxicity leading to fish mortality. The lethal main factors involved are hypoxia and the presence of ammonia. A series of experiments was carried out on juvenile rainbow trout exposed to ammonia and low concentrations of dissolved oxygen. Survival data analysis and generalized linear models were used to model the laboratory data obtained. A quantitative model of the instantaneous death probability as a function of time, concentrations of un-ionized ammonia and dissolved oxygen was obtained. The model was fitted with data obtained at a temperature of 17°C with concentrations of un-ionized ammonia between 0 and 0.5 rag/l, concentrations of dissolved oxygen between 1.7 and 2.9 mg/l and time less than 24 hours. The survival probability of rainbow trout exposed simultaneously to un-ionized ammonia and low dissolved oxygen was found to be lower than the predicted survival probability derived from the addition of individual effects of both ammonia and hypoxia. This kind of model can be used to understand the toxic impact of run-off and to draw up guidelines for the management of urban storm water. Copyright © 1996 Published by Elsevier Science Ltd Key words--storm run-off, acute toxicity, ammonia, dissolved oxygen, Oncorhynchusrnykiss, interaction, generalized Linear models, survival data INTRODUCTION that last longer. Heavy rainfalls often result in high fish mortality. As an example, 300 tons of dead fish Storm run-offs are an important source of pollution were taken out of the fiver Seine after the 31 May for aquatic ecosystems. They differ from daily rejects, 1992 storm. occurring during dry weather periods; both in quality Based on measured concentrations in rivers during and quantity. This specific urban pollution has rainfall events, and according to their observed effects various origins: sewer in-pipe sediments, the corn- on fish, the two major causes of fish mortality are position of which is similar to that of waste water hypoxia and the presence of ammonia. (high quantity of ammonia and organic matter), road Hypoxia is caused by a considerable increase in run-off (varied wastes, urban surfaces and traffic organic matter in the fiver, arising not only from the wastes) and atmosplheric pollution (aerial dry and wet run-off itself, but also from the river sediments deposits), suddenly put in suspension again by the accelerated These discharges have some specific characteristics: water flow. The degradation of organic matter burns they generate high levels of pollution in a very short the dissolved oxygen (Hvitved-Jacobsen, 1982). time and lead to cases of acute pollution. In other Moreover, in lakes or low flow rivers, oxygen pro- respects they are highly variable in time and space, duction by photosynthesis is brought down during Some studies of the physical chemistry of these rainfall events. Dissolved oxygen concentrations discharges can be found in the literature (Ellis, 1989; measured in the receiving waters are sometimes less Ellis et al., 1992) but the biological impacts are rarely than those necessary for fish survival (Davis, 1975). investigated. Less than 1 mg/l of dissolved oxygen has very often Waste water treatment techniques have been been recorded in the Seine river during summer storm improved over m,my years without taking into (Piren Seine, 1992). These concentrations are con- account the problems of rainfall events. But it sidered lethal: the 24 h-LCs0 (lethal concentration for appears in the 1970s that the pollution discharged on 50% of the individuals after 24 h of exposure), is these occasions was considerably higher than that 1.2 mg/l for Tinca tinca and 3.6 mg/I for Thymallus of treated urban run-off(Ellis, 1986; STU, 1989). This thymallus (Kra'/em and Patte¢, 1980); Shepard (1955) occasional pollution can generate acute effects that found a 96 h-LCs0 after of 1.75 mg/l for Oncorhynchus stop with the polluted run-off but also chronic effects mykiss. 211