Providing winter food for farmland birds: Who visits feeders and does it work?

02 November 2023 - Published by Francis Buner
Supplementary winter feeding with gamebird feeders was part of our PARTRIDGE management toolbox. But did our target species actually use these feeders? And do best-practice guidelines for feeder management have an effect? To answer questions like these, camera traps were used to monitor the animals visiting our feeders.

Winter weather can pose serious challenges for birds living on modern farmland. Besides an increased lack of habitat cover for protection from harsh weather and hungry predators, scarce seed-food availability, especially during the ‘hungry winter gap’ in late winter, adds to the serious risk of increased winter mortality. In addition to our habitat measures, which aimed to provide better cover, protection from raptors and a reliable source of seeds into mid-winter, we provided additional supplementary winter food through feeders at our PARTRIDGE demonstration sites, wherever it was feasible, to increase and extend seed-food availability into early spring. The aim of winter feeding is to increase the winter survival-rate of seed-eating farmland birds and help them improve their body condition for the next breeding season.

As with all measures, there are guidelines on best practices for supplementary feeding to maximise the benefit for gamebirds and songbirds. These guidelines also help to reduce the amount of food taken by pest or ‘un-welcome’ species, such as rodents, corvids (e.g. crows and magpies) and pigeons. While we strived to follow these guidelines as closely as possible, deviations in the field nevertheless occurred. Camera trapping revealed the consequences of these deviations and allowed us to adapt our feeder approach.



Figure 1 – The aim of winter feeding is to increase the winter survival-rate of seed-eating farmland birds and help them improve their body condition for the next breeding season. Left: a pair of grey partridge, Right: a yellowhammer


The time involved in monitoring wildlife visiting feeders using camera traps is not a job that should be underestimated. We therefore restricted our monitoring to three of our ten demonstration sites. We used camera trapping at two sites with no recent history of supplementary feeding, namely Balgonie (Scotland) and Oude Doorn (The Netherlands). Here we managed the feeders ourselves according to current best-practice guidelines, as two of our placement students explained in their enthusiastic vlog. Additionally, we monitored feeders in our demonstration site Ramskapelle (Flanders, Belgium), where local hunters already ran their own winter feeder management for many years. At the latter we provided in-depth advice to the hunters on how to use feeders according to the best-practice guidelines mentioned above.

Over five subsequent winters we placed camera traps at 45 feeders for a total of 1,662 monitoring days, spread roughly equally over the three sites. Analysing the nearly 344,000 photographs from these camera traps can be challenging. When looking at the different species visiting our feeders, it became clear that their feeding behaviour had an impact on the number of photographs taken and the interpretation of a single visit. For example, a species can visit the feeder only once but for a long time, or frequently for short times during a single day. This complicates the comparison of feeder use between different species. To circumvent these issues, we recorded whether a species visited the feeder during a specific monitoring day. For ease of interpretation, we grouped certain species into species groups. For example, all rats and mice were grouped as rodents.

Based on this data, we determined how likely it was that a species or group would visit a feeder during a single day, the so-called daily observation probability (see Figure 3). Species visiting our feeders frequently will have a high daily observation probability. This showed that of all our project target species (namely partridge, pheasant and songbirds), the pheasant was the one that most frequently visited the feeders, especially in Belgium and The Netherlands. Note that in both these countries, releasing pheasants is not allowed and many wild broods were recorded by bird surveyors. Songbirds, on the other hand, were using the feeders more often in Scotland, and to a lesser extent in The Netherlands and Belgium. This effect was driven by the Yellowhammer (Emberiza citrinella), which was only recorded in high numbers at our site in Balgonie (among the sites used for testing feeders), while at our Belgian and Dutch demonstration sites the yellowhammer did not occur. The absence or lower abundance of songbirds eating larger seeds, such as wheat (the main seed provided in gamebird feeders), might explain the lower observation probability of this group in the latter countries. In addition, the use of a special rat-proof cylinder instead of a spiral to distribute the food might have reduced the use of the feeders by songbirds in The Netherlands, since less food was spilled on to the ground and hence available for small birds (the rat-proof cylinder was only used in the Netherlands, while the spiral was used in Scotland and primarily in Belgium).



Figure 2 – Yellowhammers were frequently observed at feeders in Balgonie, Scotland. Also note the robin and blackbird in this picture


Somewhat disappointingly, the partridge was not spotted very often and this was consistent across all sites. This further supports our hypothesis that supplementary winter feeding may not be strictly necessary for this species during mild winters. Practical experience by hunters across Europe indicate that supplementary winter feeding can be an important measure to get partridges through freezing and snowy winter periods. However, such harsh conditions did not occur during our project (except a short-lived 3-week period with snow cover in 2021). Unsurprisingly, we were unable to detect any winter effect on their feeder use. Supplementary winter feeding may reduce post-winter dispersal and improve the body condition as was found in a study on pheasants, but these aspects were not assessed in our project.


Figure 3 – Daily observation probability per species category and country: point estimate and 95% credible intervals. During the monitoring period, there was only one observation of a roe deer in Belgium and one observation of a waterbird in Scotland. No results are presented for that category/country combination


When looking at pest or ‘un-welcome’ species, namely pigeons, rodents and corvids, there was a clear difference between Belgium and the other countries, where the levels of visits for these categories were comparable. These species occurred more often at the feeders in Belgium, especially rodents and pigeons. As mentioned previously, the feeders in Belgium were managed by the local stakeholders, while at the other sites our project partners managed the feeders. Despite our advice, feeders in Belgium were not fully managed according to the best-practice guidelines by the stakeholders. As our project aimed to demonstrate best-practise examples, we set up a small field trial in Belgium during which we managed four feeders ourselves over two winters. We monitored the use of these feeders through camera trapping, while simultaneously monitoring four nearby feeders managed by the local stakeholders (i.e. traditionally-managed feeders). Despite the small sample size of this trial, we were able to demonstrate the impact of the best-practice feeder management, namely a clear reduction of visits by rodents and, to a lesser extent, by corvids. For pigeons we could not detect a difference. There was also no difference between both types of management for pheasants (Figure 4).


Figure 4 – The number of days per observation for each species category for each management type in Belgium (site Ramskapelle): point estimate and 95% credible intervals. Significant differences are marked with an *




Figure 5 – Example of a traditional feeder managed by the stakeholders in Belgium. The spiral nozzle hangs too high for partridges and will not prevent rodents from feeding directly off the feeder. Also, this feeder is permanently fixed to one location, making it easy for rats to establish themselves nearby. This may facilitate the spreading of diseases and negatively affect the target species during the breeding season, when rats will either predate eggs or chicks of ground-nesting birds


Other species groups visiting our feeders were waterbirds, hares and rabbits, and different species of predators (mainly fox and domestic cat). In Scotland, roe deer also frequently visited our feeders. Roe deer can knock over the feeders. In these cases, the use of well-framed feeders with excluders may be considered. However, frames and wire mesh can be used by rats, climbing to better access food. We therefore opted for solid tripod feeders which can cope with some interference from roe deer.



Figure 6 – Roe deer can still knock over this solid tripod feeder


Above, we described the use of feeders during the winter months overall. Here, we now dive a bit deeper into the camera trapping data to find out if there was a pattern in the visits as the season progressed. For roe deer, hares and rabbits, there was no such difference, but for the other species we detected a trend during the season. Interestingly, rodents were the only group showing a decline in feeder use as the winter progressed. The other species (partridge, pheasant, songbirds, pigeons, corvids, waterbirds, and predators) showed the opposite trend, namely an increase in the use of the feeders over the winter months, reaching its maximum towards the end of our feeding period at the end of April. This might be explained by the increase of daylight length by the end of the period, giving diurnal species more time to visit the feeders. Alternatively, as the availability of seeds in the field diminishes, the feeders might become more important for these species as an alternative food source. This corresponds with the results of our wintering bird monitoring, which showed a decline in the use of the habitat measures providing seed resources towards the end of the winter season.

In conclusion, our camera trapping data provides insights into the use of feeders by both target and non-target species. It further supports the need to comply with the guidelines for best feeder practices as we confirmed their effectiveness in reducing the number of ‘un-welcome’ species, in particular rodents and corvids. The data also validates our recommendation that the usefulness of supplementary winter feeding by feed hoppers is most important from February until the end of April. Rodents benefited the most from feeding early in the autumn/winter and target species the most towards the end of the winter period. Overall, our results highlight the need for a wider adoption of the best-practice feeder guidelines. In situations where these best practices cannot be achieved, the negative effects of supplementary feeding might well outweigh their benefits for seed-eating birds by attracting pest species in agricultural areas, facilitating disease transmission and attracting (or even increasing) predators such as rats that are likely to have negative effects on ground-nesting birds during the subsequent breeding season (by eating their eggs or young). This in turn is likely to be counterproductive for recovering the birds that were fed during winter.



Figure 7 – In situations where the best practices cannot be achieved, the negative effects of supplementary feeding might well outweigh their benefits for seed-eating birds, for example by attracting predators


As stated earlier, supplementary winter feeding was part of our PARTRIDGE management toolbox, but whether it is the right tool very much depends on its correct implementation and the project-specific aims. Given the high use of the feeders by pheasants, winter feeding can be an important tool for small game hunters. Our results further highlight the need for a wider adoption of the best-practice feeder guidelines, especially by hunters who tend to favour traditional management practices over new best-practice guidelines. This underlines the need for good advice or improved legislation to bring about behavioural change. In the Netherlands, where supplementary feeding of small game for hunting purposes is forbidden, the risk of negative effects and the restricted positive effects for partridges and songbirds might not justify the required effort.



Figure 8 – Given the high use of the feeders by pheasants, winter feeding can be an important tool for small game hunters. Note that the spiral nozzle of this feeder is set too high for partridges and that the fixed position has created a mud puddle directly below, which is an ideal place for diseases to establish themselves. Therefore, the management of this feeder does not follow best-practise guidelines


Camera trapping is a time-consuming monitoring method, but it is also a useful tool to get insights from the field which allows the adaptation of a chosen management approach in order to maximize the effect of the implemented measures. Especially in project areas with no recent history of supplementary feeding, we recommend the use of camera trapping in order to evaluate the impact of this measure before introducing supplementary winter feeding as a management tool. By sharing our findings we hope to have provided conservation practitioners with a better understanding of the benefits and pitfalls of supplementary winter feeding using feeders. We would like to extend our gratitude to all who have helped setting up the cameras, with special thanks to all the volunteers who spent many evenings identifying the animals on the numerous photographs taken.

Written by Thomas Scheppers, INBO - Belgian PARTRIDGE partner, Fiona Torrance, GWCT - Scottish PARTRIDGE partner, Jochem Sloothaak, Brabant Landschap - Dutch PARTRIDGE partner, Miel Cnuts, Hubertus Vereniging Vlaanderen - Belgian PARTRIDGE partner.  Edited by Francis Buner, Senior Conservation Scientist at GWCT and Head of PARTRIDGE