The grey partridge Perdix perdix is an important management concern in the European farmland. Pair numbers severely declined during the 20th century. As a result, the species has been listed in SPEC category 3, i.e. ‘Unfavourable’ conservation status in Europe. The largest population of western Europe occurs in France. Its status there is, therefore, decisive for the European conservation status of the species as well as for the future of the species. Populations of partridges in central northern France have been routinely surveyed since the 1980s for hunting management purposes. In this paper, we use this long-term and wide-scale survey to portray the demographic status of partridge populations. We emphasise the amplitude of spatiotemporal variations in breeding densities. In the 2000s, a number of areas where agriculture is intensive and where the species is hunted still sustain > 50 pairs/km2, whereas densities are < 5 pairs/km2 in other areas. These low densities are, however, higher than those commonly reported from other parts of Europe. Density levels exhibit large differences at a small spatial scale and show large year-to-year fluctuations which make trend assessment difficult. The 1994–2004 and 1999–2004 trends displayed different patterns; densities increased, decreased or were stable depending upon agricultural region.
Like many bird species associated with agricultural land (e.g. Siriwardena et al. 1998, Rocamora & Yeatman-Berthelot 1999), the grey partridge Perdix perdix has experienced a dramatic decline in Europe since World War II (Tucker & Heath 1994). According to Potts (1997), breeding stocks have dropped by > 80% since the 1930s. As a result, the grey partridge has been listed in SPEC category 3, i.e. ‘Unfavourable’ conservation status in Europe, but with main numbers outside Europe (Tucker & Heath 1994, Aebischer & Kavanagh 1997). Recently, BirdLife International (2004) drew attention towards this species because “although the species was stable or increased in many eastern European countries during the period 1990–2000, it has continued to decline throughout most of western and central Europe - including sizeable populations in France and Poland - and underwent a large decline overall”. This decline is all the more worrying because agro-environmental schemes have been applied in most countries for the benefit of wildlife (see Kleijn & Sutherland 2003, Bro et al. 2004).
The European breeding stock of grey partridge (excluding the Russian and Turkish populations) was estimated at 1.7–2.9 million pairs in the mid-1980s (Aebischer & Kavanagh 1997), and at 1–2.3 million pairs in the 1990s (BirdLife International 2004) with the largest population occurring in France. Reitz (2003) estimated the breeding population size at ca 750,000 pairs in France in spring 1998. The status of the French population is, therefore, decisive for the overall European conservation status of the species as well as for the future of the species in western Europe.
In France the grey partridge is a culturally important gamebird. Within the sedentary small game species inhabiting cultivated plains, the grey partridge ranked fourth in number of individuals killed by hunting after pheasant Phasianus colchicus, rabbit Oryctolagus cuniculus and red-legged partridge Alectoris rufa (ONCFS 2000). Ca 1.5 million birds were shot during the 1998/99 hunting season (Reitz 2000), but the proportion of released birds in the bag was unknown. Tupigny (1996) estimated that two million grey partridges were reared in France in 1995. However releases are mostly practised in regions where the species has almost disappeared (Reitz 2003a). In central northern France, where the species is still well represented, progressive hunting management has been practised since the mid-1980s (see Reitz 2003a) to ensure sustainable hunting (Aebischer 1997) and to preserve wild birds. Within this context, a survey of partridge populations assessing breeding density and reproductive success is conducted annually (see Reitz 1999, Bro et al. 2003). The primary objective of this survey is to calculate bag limits (quotas) to adjust hunting pressure to pair density and reproductive output at a local scale (Reitz 2003b).
In our paper, we use this long-term and wide-scale survey to portray the status of the grey partridge in central northern France, emphasising the large spatiotemporal variability. We present a map of 2000–2002 breeding densities at the ‘commune’ scale and assess both the 1994–2004 (long-term) and 1999–2004 (recent short-term) trends in densities for each ‘farming region’.
The French national partridge survey
Monitoring of grey partridge populations was initiated in the early 1980s in a few areas and was progressively extended to many other areas in central northern France. Spring censuses and brood surveys have been conducted every year to assess breeding density and reproductive success (see Reitz 1999, Bro et al. 2003).
Because the primary objective of this survey is to calculate hunting bag limits, these areas correspond to hunting estates where hunting management is undertaken to manage wild grey partridge populations; only few of these hunting estates practised releases of a small number of hand-reared birds for shooting purposes (Fig. 1). Therefore these areas were not selected at random or using a particular sampling procedure. However, extensive counts are carried out in some farming regions, hence they correctly reflect partridge status (Fig. 2 & Table 1, and see the sections ‘Statistical analyses’ steps 3a and 3b and ‘Discussion’).
Levels of grey partridge density (pairs/km2, mean ± SE) in each farming region of central northern France in 1994–2004. Proportions of the communes and arable land sampled were estimated using data from 2000–2004. This information is given to assess data representativity. The proportion of arable land censused is reported as classes because the area actually censused may vary from year to year.
Spring censuses were carried out to estimate the breeding stock. Counts were performed in March, when birds had paired and before the crop cover was too high (in particular oilseed rape and winter cereal). We censused partridges that flushed from the sample plot while fields were beaten by a line of people (see Reitz 1999). To achieve a census as complete as possible, 20–50 people were needed to count one sample plot depending upon its area (ranging approximately within 80–250 ha). Sample plots were representative of the ‘commune’ (Fig. 3).
Spring counts were reported either as (i) the number of pairs, trios and single birds (where density levels were low) or (ii) the total number of birds. In the former case, the number of pairs was calculated as the number of pairs and trios plus the number of single birds divided by 2.1; in the latter case as the total number of birds divided by 2.1 (Reitz & Berger 1994). The number 2.1 corrects for the unbalanced sex ratio in spring (Birkan & Jacob 1988). The spring sex ratio was estimated in the field by examining pairs and single birds using binoculars.
We estimated the trend of breeding density as the regression slope of density (previously log-transformed) against year (continuous variable). We used an autoregressive error model to diagnose and correct for serial correlation due to time series (proc AUTOREG - first-order autoregressive error, maximum-likelihood method). Missing values in time series were not filled because the procedure permits embedded missing values for both the independent and dependent variables.
The statistical unit was the ‘commune’ (see Fig. 3); density level in the commune was the average of densities estimated on sample plots. Communes were pooled to estimate density trends in farming regions. Trends were estimated for both the 1994–2004 (long-term trend) and 1999–2004 (recent short-term trend) periods, using communes with at least eight and five year's data, respectively. Trend analysis was performed in three steps:
a regression was performed for each commune. Then we combined the results of these separate analyses in a file (farming region, commune, slope of the regression and standard error (SE), r-square of the model (r2), length of the time series (N) and P-value testing whether the slope was null) to be analysed in step 2;
we conducted a meta-analysis for each farming region to test whether the distribution of the slopes was significantly different from zero (proc UNIVARIATE - sign-rank test because assumptions of parametric tests were violated);
a) because counts were not carried out in random areas (see the section ‘Field procedures’) and this may bias results, we tested the robustness of the meta-analysis result using a resampling procedure. A random sample of ca 75% of communes was drawn from a uniform distribution (if ranuni(−1) ≤0.75 then selected = 1 else selected = 0) for each farming region. The seed was the computer clock. A meta-analysis (step 2) was performed on the random sample;
b) we ran all of procedure 3a) 100 times using an iterative macrovariable.
Range of breeding densities in 2000–2002
Breeding densities (of wild grey partridges) of > 50 pairs/km2, and even exceeding 70 pairs/km2, still occurred in central northern France in the early 2000s (see Fig. 2 and Table 1). However, in other areas, density levels were as low as a few pairs/km2.
The map of mean 2000–2002 breeding densities showed sharp contrasts in density levels at a small spatial scale (see Fig. 2). High density (i.e. > 50 pairs/km2) areas were not gathered in a core region but were scattered both in northern and southern regions (see Fig. 2 and Table 1).
Trend patterns were contrasting across farming regions (Table 2). As for density levels, positive and negative trends were not geographically gathered, but distributed as a mosaic. However, few trends were statistically significant, either because the sample size (number of communes sampled) was low or because trends were not obvious due to large year-to-year fluctuations in density levels (Fig. 4 and see Discussion).
Trend in grey partridge density in each farming region of central northern France. See text for explanation about statistical analyses (full meta-analysis: step 2, resampling procedure: step 3). Results of the resampling procedure are given when the sign-rank test probability of the full meta-analysis is significant and/or when the number of communes is higher than 10.
The four regions Ponthieu, Vimeu, Plateau picard and Champagne crayeuse showed an increase in partridge density in the long term. This result was driven by a recent increase in density for Plateau picard and Champagne crayeuse. Contrarily, partridge density decreased in the Perche and Beauce regions. Short-term and a long-term declines were observed in the Gâtinais and Drouais-Thymerais regions, respectively.
Density levels of grey partridge are very contrasting from one commune to another (and even at an infra scale; E. Bro, unpubl. data). Areas with densities as high as 50 or even 70 pairs/100 ha still exist in the early 2000s, whereas densities declined to a few pairs/100 ha in other areas. High and low density areas are spatially distributed as a mosaic at a small spatial scale, suggesting that high densities result from local, not global factors. However, environmental factors determining such variability is beyond the scope of this paper. Relationships between density and habitat characteristics were recently investigated at three nested spatial scales and a paper reporting results is currently in preparation.
The levels of density commonly found in France, not only high densities, are quite favourable compared to those frequently reported from other European countries (Table 3). The Polish partridge population that ranks second in size in Europe after the French population (Aebischer & Kavanagh 1997, BirdLife International 2004) sharply declined during the 1990s and densities reached an average of a few pairs/100 ha in many regions (Panek 2005). Densities as high as ca 20 pairs/km2 or even exceeding 60 pairs/km2 were recently reported as partridge ‘hotspots’ in suburbs of Praha (Salek et al. 2004) and Frankfurt (Kugelschafter & Richarz 2001). But these areas seem to be spatially restricted, and large city suburbs are not designated as long-term wildlife conservation areas.
Levels of grey partridge breeding density in some European countries and USA in the 1990s and 2000s.
Trends in numbers
Grey partridge numbers are reported to decline at the national scale. Reitz (2003a) estimated that the overall abundance of the species had decreased by ca 20% during 1979–1998. An other independent source of data in France reported a 49% decline of partridge numbers between 1989 and 2001 (CRBPO 2005).
Decline in numbers can result from an overall decline in densities or/and a range contraction. The decline of the grey partridge in France seems to be a combination of the two phenomena. This work did not allow range assessment because counts were carried out in areas where partridge densities were high enough to encourage hunters to survey and manage wild populations, but Reitz (2003a) documented this point comparing two inquiries. He reported a range contraction of the species in France between 1979 and 1998. The grey partridge has declined most in the Bretagne, Lorraine, Franche-Comté, Limousin, Auvergne and Rhônes-Alpes regions. The situation is more complex in central northern France where trend patterns were contrasting at the regional scale. Densities were roughly stable over the last decade in all but a few regions where densities either decreased (Perche, Drouais-Thymerais, Beauce) or increased (Ponthieu, Vimeu, Champagne crayeuse, Plateau picard). These findings are similar to Reitz's results using independent data sets (Reitz 2003a). Indeed, he found that the proportion of communes with densities of > 15 pairs/100 ha increased between 1979 and 1998 whereas the proportion of communes with densities of 6–15 pairs/100 ha decreased, but he observed an overall decline. These convergent results provide confidence in our conclusions (opposite trends depending upon regions). Yet the problem of low-density areas (few pairs/100 ha) persists. Indeed when partridge densities reach low levels, hunters lose their interest in the species because they cannot hunt it any longer (unless they release hand-reared birds for shooting purposes), and they often stop the survey. The other available methods to monitor partridge populations (survey using farmer's observations (Brun et al. 1990), prospection of linear features (Brun et al. 1990), presence of faeces (Pinet et al. 1981) or male call counts (Panek 1998)) are time-consuming and except the first one, they are not extensively carried out into practice. As a consequence, long-term surveys are biased towards high density areas (i.e. central northern France) and population dynamics at low density are little documented. Unfortunately, other monitoring programmes, such as the STOC (CRBPO 2005) or the ACT (Boutin et al. 2003) are not suitable for the grey partridge and thus can not complete our survey at the national scale.
We estimated trends in density performing a meta-analysis on regression slopes of log-transformed density levels against years. This procedure may nevertheless be questioned for species such as the grey partridge whose population dynamics are highly fluctuating from year to year (see Fig. 4). Our long-term survey showed that a number of areas sustained very high densities (> 70 pairs/km2) only temporarily. Peaks often occur simultaneously in a number of areas (for instance Petit Caux in 1997 and Beauce in 1991; see Fig. 4) and are generally related to a good reproductive success the previous year (E. Bro, unpubl. data). A positive or a negative trend may result from such peaks when they occur at the beginning or at the end of the survey. Siriwardena et al. (1998) recommended the use of smoothed index series to solve the problem. We preferred to describe trend patterns providing the corresponding data (see Tables 1 and 2).
Together with other farmland birds, the grey partridge has been reported to be a species whose conservation status has worsened alarmingly (see BirdLife International 2004). Indeed, numbers have continued to decline over the last decade in western Europe despite the application of agro-environmental schemes in most countries (see Kleijn & Sutherland 2003). Moreover, the species is all the more exposed because it is huntable (listed on annex II of the Bird Directive). For all that, hunting is not necessarily the reason for the persisting decline. Sustainable hunting may be justified (see Ellison 1991, Aebischer 1997). Our survey of the grey partridge in farmlands shows firstly that high density wild populations still occur in intensively cultivated areas (cereal ecosystems where wheat yield may reach 120 Q/ha) where partridge hunting is culturally important. Secondly, it highlights that except in a few farming regions, densities did not particularly decline during the last decade but fluctuated from year to year. In most of central northern France, the species is likely to benefit from sustainable hunting in three ways:
Hunting provides financial resources for scientific research (carried out by the governmental Game & Wildlife Agency, ONCFS). The population dynamics of the grey partridge are uniquely well studied among farmland birds. Research involves a large-scale and long-term survey (this paper), large-scale experiments (see Bro et al. 2004) and widespread field management (Bro et al. 2004). Understanding the cause of its decline allows scientists to make recommendation about conservation and agricultural management (see Potts 1997). Some management prescriptions were included in recent CAP reforms.
Hunting management requires annual and local field data to attribute quotas to hunting estates, hence it provides long-term and wide-scale monitoring of partridge populations.
Hunters invest time and effort in managing farmland habitats and controlling predator abundance to increase hunting bags. This land management favours partridge abundance and productivity (Tapper et al. 1996), and this often results in higher densities in areas where the partridge is hunted than in areas where the partridge is not hunted. This is what N. Aebischer named “the paradox of wise use” (Aebischer 1997). Such generic conservation actions are likely to benefit other farmland species, also.
The French ‘grey partridge national network’ is a collaboration between the Federation Nationale des Chasseurs (FNC), the Fédérations Départementales des Chasseurs (FDC) and the Office National de la Chasse et de la Faune Sauvage (ONCFS). We acknowledge all field technicians of FDC and ONCFS and hunters who have been collecting the data for many years.