Control of a coypu Myocastor coypus population in northern Italy and management implications

Abstract

The effects of removal on the structure of a coypu population were investigated during a three‐year trapping program in the Val Campotto wetlands (northern Italy). The investigation was conducted on 267 specimens collected from March to December 1996. Winter severity index values indicated a sequence of cold winters, which may explain the unusual lack of individuals in the 0–4 months age class. Nevertheless, the demographic changes due to cold winters were counterbalanced by a high birth rate, the presence of pregnant females throughout the year and a low rate of embryo resorption (suggesting a high recruitment rate). Despite the intensive trapping campaign from 1994 to 1996 (8600 animals removed), no reduction of coypus was observed. Control campaigns and cold winters temporarily affected the sex ratio and age structure of the population, but did not limit coypu abundance. These results were compared with data collected during similar coypu control programs carried out in Italy and abroad.

Keywords:

• Myocastor coypus
• pest species
• population control
• demography
• management
• population structure

Introduction

The coypu Myocastor coypus (Molina, 1782) is a South American rodent which was introduced into Italy for use in fur farms in 1928 (Capocaccia Orsini & Doria 1991). Since then, there have been several escapes from captivity, particularly in central and northern Italy (Santini 1981). Escaped animals became naturalized in large areas of the country. At present, the species occupies two large zones in northern and central Italy. The former extends eastward throughout the Po Valley to the Adriatic coast, while the latter includes the plains and hills of Tuscany and Latium between the Arno and Tiber rivers. Wetlands, marshes and slow‐flowing streams are suitable habitats for the coypu, as are reclaimed lands. In southern Italy and the major islands, the coypu distribution is still localized in several small populations separated by hundreds of kilometres (Cocchi & Riga 2002).

In recent years, ecological, hydraulic and economic problems caused by coypus have been increasing rapidly in several countries, including Italy. The ecological impact consists mainly of structural alteration of aquatic ecosystems (D'Antoni et al. 2002). Prigioni et al. (2005) found that 7 of 12 threatened aquatic plants in the Ticino regional park were eaten by coypus. Moreover, there is evidence that coypus destroy bird nests in nesting sites (e.g. Gariboldi 1993). In NE Italian wetlands, coypu control has resulted in an increase of some bird species, such as the great crested grebe Podiceps cristatus, little grebe Tachybaptus ruficollis and whiskered tern Chlidonias hybridus (Tinarelli 2002). Digging by coypus in drainage canals can affect the stability of banks and even cause severe floods with significant costs and risks to human safety (Panzacchi et al. 2007).The rodent is also responsible for extensive damage to crops and population control programmes aimed at limiting the impact are carried out by several local authorities (Italian Wildlife Institute archives). Farmers' discontent is increasing and there is frequent demand for control measures. Under Italian law, coypu removal can only be undertaken by local administrations and by staff of protected areas. The two allowed removal techniques are live trapping and direct shooting.

The aims of this study were: (a) to assess the effectiveness of a removal program carried out in Val Campotto a wetland in the eastern part of the Po river Delta (NE Italy), (b) to study the coypu population demography, and (c) to evaluate the relationship between winter severity and reproductive parameters.

Materials and methods

The study was carried out in Val Campotto, a 9.6 km² semi‐aquatic wetland ( 44°35′N, 11°47′E). Val Campotto is an artificial embankment created to collect excess water flowing down from the Apennine Mountains reducing, in this way, potential hazard because of sudden seasonal fluctuations. Campotto is less than 30 km from the Adriatic Sea. The topography is flat with an altitude ranging between 1 and 4 m above sea level. In 1996 the mean daily temperature was 12.5°C with a maximum of 22.4°C in July and a minimum of 2.5°C in February. Average annual precipitation was 84.5 cm, with peaks in December and October. During winter 1995–96 there were 48 days with temperature equal to or below 0°C. Reed (Phragmites australis and Typha latifolia) thickets are frequent in shallow waters, whilst water lily Nymphaea alba is the most widespread species in deeper waters (to 1.5 m); woods of ash Fraxinus oxycarpa associated with white poplar Populus alba and elm Ulmus minor are an important relict of Padania hydrophilic woods, whilst embankments have been reforested as windbreak barriers. Val Campotto is part of a larger semi‐aquatic ecosystem including important areas such as the Po Delta, Comacchio marshes and a network of relict coastal wetlands. Because of the variety and the abundance of its biocenosys Campotto is a protected area (Ramsar zone), where hunting is forbidden.

Between January 1994 and December 1996, a coypu trapping programme was carried out in a portion of the Val Campotto (7.5 km²). The control schedule was heterogeneous, being interrupted by several periods of inactivity (8 of 35 months without control). During the three years of trapping, there was an alternation of low trapping (March–May 1994, Dec 1994–Jan 1995, Nov 1995–Jan 1996) and high trapping effort (maximum of 2200 trap‐nights in June 1995). From March to December 1996 we analysed a sample of 267 specimens captured with double‐entrance live cage traps placed on the ground and on floating rafts. All animals were shot immediately after capture. For each individual were recorded weight (W), sex, total length (TL), head‐body length (HBL), foot length (FL) and reproductive status. Sex was determined by observation of the external genitalia (Ehrlich 1966). Sexual maturity was estimated by external observations (vulva perforation in females and development of the penis and presence of spermatozoa in males). Reproductive tracts of females were inspected during dissection to detect implanted embryos, foetuses reabsorbed and to assess uterine development. Age was calculated from the eye lens dry weight according to the equation of Gosling et al. (1980):

where 4.34 is the coypu gestation period in months. On the basis of the individuals' age at capture we assessed their birth time. Assuming that higher is the weight/body length ratio better is coypu physiological condition, we calculated the condition index (K) using the method described by Bailey (1968) for Eastern cottontail Sylvilagus floridanus:

where W is the body weight in grams and HBL is the head–body length in decimetres. To age embryos (in days) we used a linear relationship with W³ (1 mg accuracy) following Newson (1966) and an updated conversion list for embryo age estimation from weight (L.M. Gosling, personal communication). For each pregnant female the number of foetuses and their age (in weeks) were determined. Furthermore, we discriminated embryos by weight. Embryos heavier than 58 g (type B) were assumed to be older than 14 weeks and less prone to reabsorption (Gosling 1986a, 1986b). We calculated the winter severity index (CRS) following Gosling (1981b) as modified by Reggiani et al. (1993):

where x is the length of a run of consecutive freezing nights (minimum temperature <0°C) and n is the number of runs each winter. The variant introduced by Reggiani et al. (1993) appears more suitable because it allows the inclusion of days where freezing nights are followed by midday temperatures higher than 5°C. Two aspects must be considered in this regard. Firstly, coypus concentrate social and feeding activities at night (Gosling 1979), but reduce nocturnal activity in very cold weather when energetic costs exceed the benefits of feeding intake (Gosling 1981a). Secondly, at the latitude of our study area, cold nights can occur at the beginning and end of winter. Hence, we considered the nocturnal minimum temperatures throughout the season (from November until March). Three consecutive CRS indexes were calculated. The first referred to the winter before the capture period and the others to the two previous winters. The F test was used to assess morphological differences between the sexes, whilst Spearman's rank correlation test (rs) was used to analyse the relationship between winter severity and birth distribution. Removal effectiveness was evaluated by monitoring the trapping success (coypus trapped/100 trap nights) at monthly intervals. This index was related to the cumulative trapping effort (number of trap nights in the capture period March–December 1996). A logarithmic regression model was used to test the relationship between the condition index K and the temperature and between catching success and cumulative trapping effort (Velatta & Ragni 1991).

Results

The only significant morphometric difference between males and females was for foot length (F _1,210 = 7.468, P = 0.007).

Births occurred all year long with a peak between June and August (41.6%) and the lowest values in December (3.7%) and January (4.1%). 74.5% of mature females and 58.2% of all females were visibly pregnant. Apart from a period of missing data (from January to February 1996), pregnant females constantly represented a high percentage of the mature ones (monthly percentage higher than 60%; Figure 1).

Figure 1. Percentage of pregnant females at Campotto.

In December the only mature female found was not pregnant. The mean litter sizes calculated using both foetuses of type A (n = 370,  = 5.52±1.73) and type B (n = 76,  = 5.06±1.91) did not differ significantly (t = 1.926, df = 80, P = 0.7585). In the period March–December 1996, 8.74% of embryos were reabsorbed. In type B foetuses we found more males than females (1M:0.87F; χ²  = 5, P<0.05).

The average age of the trapped animals was 13.8±8.16 months. Males reached sexual maturity within 6–7 months of age, females at 3–4 months. The sex ratio for all captured coypus was 1M:1.12F and for adults (n = 212) 1M:1.09F, although the differences were not statistically significant (χ² = 0.226, P = 0.359).

During the three‐year control campaign (1994–1996), overall 8592 coypus were removed from the Val Campotto area involving a total of 22,849 trap nights.

The capture curve in the study area showed an early decrease followed by a stabilisation of capture success until the end of the trial (y = 0.0003x+31.098, R²  = 0.3796).

There was no significant correlation between trapping success and mean temperature (F _1,492 = 3.052, P = 0.081). No coypu showed frozen feet while animals with shortened tails due to frost damage (Willner et al. 1979) constituted 18.3% of the total population. In the overall sample, the condition index (K) averaged 29.73±4.23 with no significant sex difference in mature individuals (F _1,210 = 0.027, P = 0.869, n = 210). The Campotto K index values showed a slight negative correlation with mean temperature, although it was not statistically significant for the whole population (y = −0.7163x+37.662, R²  = 0.0361) or for the different sexes (y = −0.4238x+28.95, R²  = 0.029 for males and y = −0.1556x+20.949, R²  = 0.002 for females). We found a positive correlation between mean monthly temperature and birth distribution (rs = 0.779, P = 0.001).

Discussion

Two factors seem to have limited the effectiveness of coypu removal program carried out in Val Campotto: (i) the inadequate capture effort and (ii) habitat characteristics that do not allow prevention of immigration from neighbouring areas (Gosling et al. 1988; Doncaster & Micol 1989; Reggiani et al. 1995). Regarding the former aspect, it must be considered that the control effort in the Campotto area was characterized by a low cage trap density (1632 trap nights/km²) if compared with that performed at Lake Trasimeno (5916 trap nights/km²) (Velatta & Ragni 1991), and by a discontinuous activation of the traps (trapping suspended for 8 of the 35 months). As concerns habitat characteristics, Campotto is part of a system of coastal wetlands, swamps and ponds connected by a dense network of drainage canals and rivers through which coypus can readily disperse. The situation at Lake Trasimeno is quite different, with a more isolated basin where the effects of the control campaign can likely persist for a longer time. In fact, a logarithmic regression analysis of the captures in the Val Campotto and Lake Trasimeno areas showed a substantial difference. At Lake Trasimeno it was estimated that captures would be reduced to zero at 64,700 trap nights (rs = −0.819, P<0.01; Figure 2B), whereas at Campotto the cumulative trend of coypu captures did not predict total removal (Figure 2A).

Figure 2. Efficiency of two control programmes: (A) Campotto, (B) Trasimeno.

The number of trapped coypus was inversely correlated with cumulated trap nights only in the Trasimeno study area (rs = −0.819, P<0.01). At Campotto instead, the trend was too slight to achieve a zero‐capture point in a relatively short period. It was estimated (via regression analysis) that zero capture would be reached only after thirteen years.

Bertolino et al. (2005) also reported a different effectiveness of coypu control in two areas with different ecological and hydraulic characteristics. In a small isolated natural reserve, a 3‐week control campaign led to the removal of 90% of the population with limited recolonization in the following years. In a nearby canal, 66 coypus were trapped, but there was no significant reduction in trapping success and other animals coming from adjacent territories reoccupied the removal area in the following year. These data suggest that coypu eradication is far more practicable in isolated and small areas with low population densities. On the contrary, the results of our study confirm that eradication of the coypu from continuous optimal habitats requires intense trapping efforts, carefully calibrated to population density and to the size of the area (Gosling et al. 1988; Reggiani et al. 1993).

The higher proportion of females captured at Campotto may be related to the long trapping period carried out until 1996 which may have produced a sex ratio skewed towards females as observed by Gosling and Baker (1989). The prevalence of males among the type B foetuses seems to confirm the findings of Gosling (1986b) about coypus and, in general, accords with the main pattern of polygynous mammals (Clutton‐Brock & Albon 1982). After birth, males are more exposed to mortality risks than females (Greenwood 1980) and this can explain the higher proportion of females found in the older age classes.

In agreement with the results of Velatta and Ragni (1991) at Lake Trasimeno, we did not find any correlation between the condition index and environmental temperatures. This contrasts with data collected in Maryland marshes where the K index was positively correlated with local temperature and with frost‐free days (Willner et al. 1979). Winter harshness causes frostbite injuries to coypu feet and tails. In Maryland, 25.6% of coypus had damaged tails versus only 2.0% in Louisiana (Adams 1956). At Campotto, shortened tails due to frost damage occurred in 18.3% of the population.

When compared with the Trasimeno population (Velatta & Ragni 1991), the coypus in Campotto got older (proportion of coypus >2 years: Campotto = 7.6%; Trasimeno = 1%; Figure 3). On the other hand, there was a large reduction in the 0–4 months age class in Campotto (5.2% versus 30.5%).

Figure 3. Age structure in two Italian coypu populations.

Newson (1966) in England, Willner et al. (1979) in Maryland and Reggiani et al. (1995) in the Tevere‐Farfa Reserve (central Italy) found that coypu reproduction was hindered or interrupted by cold weather. At Campotto the CRS winter severity index values were always at least one order of magnitude higher than those collected at Trasimeno (Table I). Similar differences were recorded for the number of frozen days.

Table I. CRS index and number of frozen days at the Campotto and Trasimeno study areas.

	Campotto		Trasimeno
	CRS	No. frozen days	CRS	No. frozen days
Winter before coypu capture	206	54	32	12
1 winter	459	59	39	14
2 winters	789	78	7	5

At Campotto the harsh winter of 1995–96 (CRS value of 206) may have caused the collapse of the 0–4 months age class recorded in the following breeding season. Similarly, the lack of the immature class in Tevere‐Farfa recorded in the spring trapping session may have been caused by the interruption of births following the previous cold winter (CRS = 318). This hypothesis is supported by Gosling (1981b), who found an inverse correlation between winter severity and the percentage of female littering in the following spring. On the other hand, the frequency of intermediate age classes at Campotto was similar to that recorded at Trasimeno, suggesting that the coypu is able to quickly recover its natural demographic structure after a series of climatic perturbations. At Campotto was observed a significant presence of pregnant females from March until November 1996.

Data here presented depict a demographic situation in which the effects of removal campaigns and cold winters can temporary affect coypu sex and age classes but do not influence its abundance as observed elsewhere (Gosling et al. 1981). High birth rate, presence of pregnant females throughout the year and a low rate of embryo reabsorption confirm the plasticity of the coypu and suggest the presence of a high potential recruitment, allowing the recovery of losses due to cold winters and removal campaigns. In particular, our sample did not show a significant difference in litter size between type A and type B foetuses unlike what was found in England (Newson 1966) and at Lake Trasimeno (Velatta & Ragni 1991).

The results of this study underline the difficulties in carrying out effective coypu removal programmes in wide areas with the presence of a continuous optimal habitat for the species. For this reason in Italy the complete eradication of coypu is an unrealistic goal. Notwithstanding this, eradication on a local scale is still feasible. In order to evaluate the eradication feasibility of potential areas, the relationship between population density, trapping effort and demographic response could be inferred by a modelling approach tested in the most common coypu habitats. This should be a key step in the development of a national strategy for coypu management defining where the eradication of isolated coypu populations is still practicable (e.g. in Sardinia and Sicily).

Acknowledgements

The study was carried out within the LIFE contract no. B4‐3200‐94‐778 funded by the European Community and Regione Emilia‐Romagna. We are grateful to Paolo Pini, director of the Consorzio della Bonifica Renana, and to his staff for their support of the project. We also thank Giovanni Vecchi, director of the Istituto Zooprofilattico Sperimentale (Bologna section), Valter Trocchi and Emilio Lenzo for their valuable contributions. Special thanks to Piero Genovesi, Dario Capizzi, Andrea Verardi and two anonymous referees for providing valuable criticisms on a previous draft of the manuscript.