Temperature and massive fish deaths in southern South America

Based on a compilation of data on massive fish deaths occurred in southern South America during the Twentieth Century, we assessed the importance of climatic variables on these phenomena. We found a strong relationship (R2 =0.68) between these massive fish deaths and the mean monthly air temperature. Along the annual temperature range there is a central range (14.6° C to 20.0° C), where the probabilities that a fish community suffers massive deaths is very low. Its central point (17.3° C) is very close to the mean annual value (17.2° C) of air temperature variation. We considered this agreement as corroboration at community level of Pianka’s theory on physiological optima. This relationship allows to monitoring the influence of climate changes, because the environmental variation and the zones of mortality and no mortality will change with predicted changes of the mean monthly values of air temperature.


_____________
MFD have a random distribution with respect to rainfall (r=0.2795) and pressure (r=0.1338),but they have an evident relationship (Pearson's coefficient) with the mean monthly air temperature (MMT) observations in spring and summer (r=0.6056,p<0.05), and in autumn and winter (r=-0.5801,p<0.05).
MFD are more frequent in summer and winter.If the same data are represented as the absolute frequency of MFD in relation with the mean monthly air temperature (MMT) corresponding to the month of occurrence along to 95 years of observations (Fig. 2), the data fit a bell-shaped curve that shows the occurrence of more than one MFD within the range below 14.5° C and over 20.1° C.This curve has a high determination coefficient (R 2 =0.68), with more than 65% of the variance of the number of MFD explained by the variation of MMT.
A zone of no mortality (NMZ) corresponds to the temperature ranging from 14.6° C to 20.0° C, with a central point at 17.3° C.This point is a physiological optimum (PO), where species have a maximum of efficiency and where the probability of suffering a MFD due to temperature is very low.
For the bulk of Paranensean and Pampasic fis-gion (sensu Ringuelet, 1961).Fish species living in the area and susceptible to MFD are considered Paranoplatensean or Pampasic (Ringuelet, 1975), and their distribution and general ecology are discussed in Menni (2004).Ecological theory establishes that "for each environmental variable there is a physiological optimum, where the biological efficiency of the organism is a maximum" (Pianka, 1983).This statement may be conceptually derived from principles referred to fishes under laboratory conditions (Fry, 1971) and it explains numerous situations.Notwithstanding, its application to concrete data is difficult, because of the different statistical distribution of environmental variables, and because the wide concept of "biological efficiency" only may be interpreted, for the time being, as survival, growth or reproduction.
Here we examine MFD as related to the Pianka's statement on optimal values, obtaining a quantitative precise generalization related to environmental change.
For this analysis we used climatic statistics from 1961 to 1970, which is the first decade previous to the detection of global climatic change in Argentina (IPCC, 2002;Barros et al., 2006).Data were obtained from the meteorological station Aero Buenos Aires, placed at 6 m. above sea level at 34° 34' S and 58° 25' W (SMN, 1985).This station is representative of the climate variation of the Paranense and Pampasic areas in the southern parts of the Guayano-Brazilian subregion.

RESULTS
MFD have been registered throughout the year except April.The relationships between the number of monthly MFD and the monthly oscillation of temperature (°C), rainfall (mm) and atmospheric pressure (mb) (Fig. 1) show that  hes, the thermal tolerance zone (TTZ) measured in laboratory has a range between 1° C to 38° C (Gómez, 1996).The environmental variation (EV) is less, ranging from 8.0° C to 28.4° C (SMN, 1985), with a mean annual value of 17.2° C.

DISCUSSION
The NMZ here calculated based on field data (14.6°C to 20.0° C) is included within the TTZ measured in laboratory (1° C to 38° C).The difference between these intervals (1.0° C to 14.6° C and 20.0° C to 38° C) is a range where temperature is an important factor associated with death.For example, a temperature of 8° C is not directly lethal, but fishes become weak and developed pathologies due to chronic exposure to sublethal levels, which brings them to death.At the other extreme, a non lethal temperature of 32° C increases metabolism and accelerates oxygenconsuming degradation of organic matter, finally causing fish death by asphyxia.
Usually, the optimal values of a variable for a given species are obtained from individuals under experimental conditions.Here, we use an indirect reasoning based on the range where massive deaths occur, implying communities or part of them, to validate and to extend Pianka's theory.MFD and temperature data are, in this case, independent in space and time.
The calculated physiological optimum of temperature (PO=17.3°C) does not present significant differences (p<0.005,two-tailed T test; t=-0.019795;p=0.978648) with the mean annual value of the environmental variation of air temperature (17.2°C) of the environments inhabited by the considered species in the studied area (Fig. 3).During the last 40 years in the Salado Basin in Argentina (36°S), the mean minimum annual temperature have risen from 8.9° C to 10.1° C and the total annual rainfall have increased from 700 to 950 mm (Gómez & Menni, 2005).These changes have resulted in the formation of new fish communities in water bodies formed in previously dry areas (Gómez et al., 2004).
Temperature increase attributed to global change is considered one of the most important factors in producing changes in freshwater systems (Meisner & Shuter, 1992;Gooseff et al., 2005;Ficke et al., 2007).Other factors such as the available amount of water, oxygen diminution and dryness should also be considered.
The generalization proposed here may be used to predict MFD and to assess the impact of climate change on freshwater fish communities, since the values of the NMZ and the EV will change along with changes of the mean monthly air temperature.

Fig. 1 .
Fig. 1.Relationship between massive fish deaths (MFD) and climatic variables along the annual cycle during 95 years.Rainfall in mm, air temperature in °C and pressure in mb -1000.

Fig. 2 .
Fig. 2. Relationship between number of monthly massive fish deaths (MFD) (dotted line) and the mean monthly air temperature along 95 years.

Fig. 3 .
Fig. 3. Thermal zones: Relationship between monthly number of MFD and mean monthly temperatures (see Fig. 2).TTZ: thermal tolerance zone (1° C to 38° C).EVZ: environmental variation zone of air temperature (18° C to 28° C, the white triangle is the mean value 17.2° C).NMZ: non mortality zone (14.6°C to 20° C, the triangle is the physiological optimum 17.3° C = PO).Grey zones: temperature values associated to death (1° C to 14.6° C and 20° C to 38° C).