General enquiries on this form should be made to: Defra, Procurements and Contracts Division (Science R&D Team) Telephone No. 0207 238 5734 E-mail: [email protected]

SID 5 Research Project Final Report

Note In line with the Freedom of Information Act 2000, Defra aims to place the results Project identification of its completed research projects in the public domain wherever possible. The 1. Defra Project code WC0732 SID 5 (Research Project Final Report) is designed to capture the information on 2. Project title the results and outputs of Defra-funded research in a format that is easily Impact of ring-necked parakeets on native birds publishable through the Defra website. A SID 5 must be completed for all projects. • This form is in Word format and the boxes may be expanded or reduced, as 3. Contractor appropriate. The Food and Environment Research organisation(s) Agency (Fera)


ACCESS TO INFORMATION Sand Hutton The information collected on this form will York be stored electronically and may be sent YO41 1LZ to any part of Defra, or to individual researchers or organisations outside British Trust for Ornithology Defra for the purposes of reviewing the The Nunnery project. Defra may also disclose the Thetford information to any outside organisation Norfolk acting as an agent authorised by Defra to IP24 2PU process final research reports on its behalf. Defra intends to publish this form

on its website, unless there are strong 54. Total Defra project costs £ 147,148 reasons not to, which fully comply with (agreed fixed price) exemptions under the Environmental

Information Regulations or the Freedom 5. Project: start date ...... 16/02/2009 of Information Act 2000.

Defra may be required to release information, including personal data and end date ...... 30/11/2010 commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

SID 5 (Rev. 07/10) Page 1 of 37 6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...... YES √ NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow. Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer. In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000. (b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary 7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

SID 5 (Rev. 07/10) Page 2 of 37 Introduction The ring-necked or rose-ringed parakeet Psittacula kramerii is native to India and sub-Saharan Africa. It is the most widely introduced parrot species in the world and has successfully established breeding populations in 35 countries across five continents. In the UK, the main populations are currently concentrated in SE England and largely restricted to urban and semi-urban sites. The population, however, is undergoing a period of rapid growth that is likely to continue into the foreseeable future – abundance increased by over 600% between 1995 and 2007.

The ring-necked parakeet is a secondary cavity nester, occupying cavities that are either natural or have been excavated by other species (primary cavity nesters). There is evidence that the parakeet competes with other cavity nesters, both in its native and introduced range - in Belgium, there is some evidence for such competition with nuthatches Sitta europaea . In the UK, parakeets initiate nesting much earlier than native species and have relatively long incubation and nestling periods; cavities, therefore, may already be occupied when native species initiate their own breeding cycle.

Aims The overall aim was to investigate the potential impact of ring-necked parakeets on native birds. The resource for which competition is likely to be most intense is nest sites and the study, therefore, focussed on the potential impact of ring-necked parakeets on native cavity-nesting species. The study incorporated three elements: (i) analysis of existing BTO datasets concerning avian abundance, (ii) fieldwork investigating abundance of parakeets and native cavity nesters in relation to habitat characteristics, nest site selection and breeding success, and (iii) spatial modelling of potential parakeet range expansion.

Species abundance Analysis of existing data from BTO/JNCC/RSPB Breeding Bird Survey (BBS), BTO Garden Birdwatch, BTO London Bird Project and BTO/RSPB/BWI/SOC BirdTrack found no significant relationship between numbers of parakeets and those of eight native cavity nesting species: blue tit, great tit, green woodpecker, great-spotted woodpecker, jackdaw, nuthatch, starling and stock dove. Further analysis of BBS data also showed no evidence that the population growth rate of any of the native species was related to parakeet numbers.

While no significant negative impact of parakeets on native species was identified at the national scale, this analysis could not exclude the possibility that parakeets may have had a localised impact at a subset of sites where the intensity of competition was greatest. Between mid-February and mid-April 2010, therefore, three survey visits were conducted at each of 20 study sites located in London parkland to record numbers of parakeets and native species. The proportional cover of seven habitat classes (low, medium and high density woodland, improved grassland, scrub, open water and human habitation) was also recorded, along with the number of veteran trees (>5m circumference) at each site to obtain a measure of nesting cavity availability.

Analyses of these data indicated no overall negative relationships between parakeet numbers and numbers of native species, even when the potential availability of cavities had been accounted for. Analysis of the interactions between parakeet abundance and habitat, however, showed that a decrease in the numbers of green woodpecker and stock dove with increased numbers of parakeet was apparent with decreasing proportion of continuous closed canopy woodland. This result suggests that some degree of competition may exist between parakeets and these two species at sites where cavities are limited.

Nest habitat Throughout the 2010 breeding season (February to July), intensive searches of 15 of the 20 study sites were conducted regularly to locate the nest sites of all target species, of which five (ring-necked parakeet, nuthatch, starling, great-spotted woodpecker and jackdaw) were found in sufficient number to permit analysis. At each nest site, habitat characteristics were measured at three spatial levels: (i) the nest cavity, (ii) the nest tree, and (iii) the habitat immediately surrounding the tree.

There was much similarity in nest habitat characteristics of some target species, with nuthatch and starling showing the greatest overlap with parakeets. However, when comparing the nest site habitat features of individual native species between sites with zero/low parakeet abundance and sites with high parakeet abundance there was an almost complete absence of any significant differences. This is consistent with a lack of evidence for the displacement of native species into less preferred cavities.

Breeding parameters For all nest cavities that were accessible, data on laying dates and numbers of fledglings produced was collected during multiple visits using a pole-mounted miniature video camera. Considering all study sites, ring-necked parakeets initiated egg-laying significantly earlier (18-45 days per species) than native cavity- nesters - median first egg date = 24 th March. Parakeet productivity was estimated to be 2.2 ± 0.23 (n=39)

SID 5 (Rev. 07/10) Page 3 of 37 fledglings per breeding pair and 2.6 ± 0.20 (n=33) fledglings per successful pair. Mean fledgling production was higher than recorded in previous studies. Nest failure, due to nest abandonment, predation or chick death was 11.3% (7/62 nests) across all sites.

Starling laying dates were significantly later in sites with high abundance of parakeets compared to sites with zero/low abundance, but there was no significant difference in the number of young fledged between study sites with zero/low parakeet abundance and sites with high parakeet abundance for any of the four native species (nuthatch, starling, great-spotted woodpecker and jackdaw).

Current impacts This fieldwork demonstrates the potential for nest site competition in the UK as parakeets favour similar types of cavity to some native species but occupy them significantly earlier. Despite this finding, there was no evidence of a negative impact of parakeets on species abundance or population growth rate at a national scale for any native species. Furthermore, there was no evidence for any overall detrimental effect on the breeding success of any native species at the local-scale.

Cavity abundance A series of studies in Belgium have reported that nuthatch numbers were lower in areas with higher numbers of parakeets. One explanation for the absence of such a relationship in the UK might be the higher density of available cavities. The median density of cavities (entrance diameter >4cm) across the 15 study sites for which habitat data was collected was 55 cavities per ha (range: 22-94 per ha), much higher than that recorded in Belgium (8-18 cavities per ha). There was some evidence to suggest that in less wooded areas, where cavities are likely to be limited, parakeets may negatively impact on populations of stock dove and green woodpecker.

Overview Taken together, these analyses may indicate that the limiting resource, cavity abundance, is high enough to support both parakeets and native species, without any overall significant detrimental effects of parakeet on native species. It is stressed, however, that the general absence of an effect of parakeet abundance on the numbers of native species reflects the current situation and does not exclude the possibility of conflict in the future if parakeets increase in number or expand their range into habitats where the relative abundance of native species is higher and/or the availability of nest cavities is lower.

Spatial modelling – population spread Spatial modelling estimated rates of spread of parakeets out of London of 1.3km year -1, 2.9km year -1 and 4.5km year -1, representing a ‘slow’, ‘average’ and ‘fast’ rate of spread respectively. Native species differed in the rate of exposure of their national populations to parakeet population expansion over the next 30 years. Under the ‘average’ rate of spread, the species that will become most exposed are green woodpecker and lesser-spotted woodpecker. After only ten years, modelling indicated that over 30% of each of these two species’ population ranges will be occupied by parakeets.

Scope of the study The scope of the study was limited by the nature of the datasets available and the duration of the fieldwork being constrained to a single breeding season. The selection of study sites of a single habitat type (parkland) was driven primarily by the distribution of nuthatch, the only species for which a negative effect of parakeet abundance was identified in Belgium. However, although the study focussed on one habitat type this encompassed a range of parakeet densities including high density parakeet areas in which interactions with native species would be expected to first occur. Sample sizes were small in a number of parameters under investigation, particularly breeding parameters and results should be interpreted accordingly. Cavity abundance was very high and possibly mitigated any negative effects of interactions. The study was not able to address the potential existence of current detrimental impacts in sites with a lower abundance of cavities. Spatial modelling was based on the best available data, but interpretation should be done with care as there is limited information on parakeet habitat use and spatial dynamics.

Gaps in knowledge A number of gaps in knowledge were identified for which further study is recommended: (i) investigation of the dynamics of parakeet-native species interactions in areas where tree cavities are more limited, (ii) validation of potential newly established parakeet populations indicated by the spatial modelling through systematic surveys, (iii) a cavity blocking experiment to investigate the response of parakeets and native species to reduced nest-site density, (iv) natural experiments to monitor the abundance, cavity-occupancy and breeding efforts of parakeets and native species in sites at the edge of the parakeets’ expanding range, (v) refine the spatial modelling through, e.g. validation of parakeet activity thresholds with systematic surveys in areas with a range of parakeet colonisation history, (vi) investigate habitat use and spatial dynamics to help determine drivers of parakeet spread and facilitate further refinement of the spatial modelling, and (vii) collaborative studies with the ring-necked parakeet research team in Belgium.

SID 5 (Rev. 07/10) Page 4 of 37

Project Report to Defra 8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include:


the scientific objectives as set out in the contract;


the extent to which the objectives set out in the contract have been met;


details of methods used and the results obtained, including statistical analysis (if appropriate);


a discussion of the results and their reliability;


the main implications of the findings;


possible future work; and


any action resulting from the research (e.g. IP, Knowledge Transfer).

SID 5 (Rev. 07/10) Page 5 of 37 1. Introduction The ring-necked (or rose-ringed) parakeet Psittacula krameri is native to Africa and Asia but following widespread introductions has established feral populations in at least 35 countries on five continents (Butler 2003). The main European populations are found in the UK (SE England), Belgium, the Netherlands and Germany. In England, ring- necked parakeets are undergoing a period of rapid population growth, which is amongst the highest observed in British bird populations over recent years; this rapid population expansion is likely to continue into the foreseeable future (Butler 2003). The present population (winter 2010) is estimated at 30,000 birds (A. Lord, pers. comm.).

The two principal resources that ring-necked parakeets and native bird species could compete for are food and nest- cavities. Whilst parakeets have a very generalist diet across their native range, consisting of cereals, weed and tree seeds, fruit, nuts and flowers (Juniper & Parr 1998), little quantitative information is available regarding the diet of ring- necked parakeets in the UK. The population is largely confined to urban/suburban areas, where individuals are frequently observed taking supplementary food provided in gardens and the fruits and seeds on garden shrubs and trees (Cramp 1985). Due to the scale of such provisioning, it is unlikely that competition for food would be the principal mechanism for competition between parakeets and native species. Furthermore, the one native species for which a negative impact of parakeets has previously been identified, the nuthatch (Strubbe & Matthysen 2007), displays relatively little dietary overlap with ring-necked parakeet (Cramp & Perrins 1993).

Competition for nest sites is likely to be more intense. The ring-necked parakeet is a secondary cavity nester, occupying cavities that are either naturally formed or have been excavated by other species. In its native range the parakeet is said to compete with other cavity nesters such as mynas, hoopoes, roller and owlets (Lambda 1966 and Sarwar et al . 1989 cited in Strubbe & Matthysen, 2007). There is, therefore, also the potential for competition for nest sites between parakeets and native species in its European range. In Belgium, recent studies have revealed some evidence for such competition, with a negative correlation displayed between the abundance of ring-necked parakeets and nuthatches Sitta europaea (Strubbe & Matthysen, 2007, 2009). In the UK, empirical data on potential competition between parakeets and native hole-nesters is unavailable, as little work has been conducted in this area. There is, however, much evidence to suggest that availability of nest cavities, a declining resource in many areas of the UK due to the loss of old buildings and trees, can limit the population size of secondary cavity nesters (Newton 1994, Strubbe & Matthysen 2007). Parakeets nest much earlier than native species, often commencing in January, and, due to the relatively long incubation and nestling periods ( c.2.5 months from laying to fledging, Cramp 1985), cavities may already be occupied when native species are prospecting for nesting sites and commencing breeding. The UK population is currently limited in its distribution but impacts within the present range may still be detectable. Any such impacts have the potential to become more widespread in the event of the continued growth and spread of the parakeet population.

This project, therefore, focussed on the interactions between parakeets and a number of native secondary cavity- nesting species. Nuthatch and starling Sturnus vulgaris display the greatest overlap with parakeets in terms of typical nest cavity characteristics (Strubbe & Matthysen 2007), and these, therefore, were the key focal species for study. Potential impacts on other species, such as stock dove Columba oenas , jackdaw Corvus monedula , blue tit Cyanistes caeruleus and great tit Parus major , all secondary cavity nesters, were also investigated, as were those on the two primary cavity nesting species that are abundant in the study area, great-spotted woodpecker Dendrocopos major and green woodpecker Picus viridis .

2. Aim The overall aim of the study was to investigate the potential impact of ring-necked parakeets on native birds. As discussed above, the resource for which competition is likely to be most intense is nest sites. The study, therefore, focussed on the potential impact of ring-necked parakeets on native cavity-nesting species.

Specific objectives were, to: i) Identify the impact of ring-necked parakeets on the distribution and abundance of native species. ii) Quantify the likely significance of these impacts on local populations of native species. iii) Assess the likely risk to the population status of native species under different scenarios for spread of the ring- necked parakeet.

3. Approach The study incorporated three elements: (i) analysis of existing BTO datasets concerning avian abundance, (ii) fieldwork investigating abundance of parakeets and native cavity nesters in relation to fine scale habitat characteristics, nest site selection and breeding success, and (iii) spatial modelling of potential parakeet range expansion.

The former analysis aimed to determine whether there is already evidence for a relationship between ring-necked parakeets and the abundance of a range of cavity-nesting birds across its existing range, while data collected during

SID 5 (Rev. 07/10) Page 6 of 37 the fieldwork element aimed to establish the nature and effect of any competitive interaction between parakeets and native species at the local scale. The spatial modelling element aimed to investigate the potential effect of any such competitive interactions on national populations as a consequence of likely future changes in the population range of ring-necked parakeets.

The project involved fieldwork in two breeding seasons (2009 and 2010). During the first project year (2009), analysis of existing BTO datasets was conducted to investigate relationships between parakeets and native species and to select suitable study sites for intensive fieldwork in the second (2010) breeding season. Fieldwork in the first season (2009) was restricted to piloting the various fieldwork protocols and techniques. In the second season a full season of fieldwork (2010), involving surveys of birds and habitat and monitoring of nesting efforts was carried out.

4. Analysis of existing BTO datasets: Assessment of evidence for the negative impact of ring-necked parakeet on population size and growth rate using extensive BTO datasets

4.1 BBS data analysis Existing abundance data collected by volunteer participants in the BTO/JNCC/RSPB Breeding Bird Survey (BBS) was analysed to evaluate the evidence for a population-level impact of ring-necked parakeets on native cavity-nesting bird species. The data were collected between 1994 and 2008, a period that witnessed major population growth and expansion of the parakeet population. BBS sampling is based on a formal sampling framework where 1-km squares are randomly selected from those in the National Grid according to a stratified random design, with approximately 2,500 squares surveyed each year. Two visits are made to each square during the breeding season (between April and July) with all bird species counted along two parallel 1-km transects.

For the purposes of the current project, analyses were performed on a subset of data from 180 squares in the Greater London area in which parakeets were recorded in at least one of the survey years between 1994-2008. Sufficient data were available to assess changes in abundance of eight native cavity nesting species; green woodpecker, great spotted woodpecker, nuthatch, great tit, blue tit, jackdaw, stock dove and starling. Another potential competitor, lesser spotted woodpecker Dendrocopos minor , was only recorded on 11 BBS squares and considered too scarce to include in these analyses.

Two sets of analyses were carried out on the BBS dataset, the first of which looked at the relationship between the number of ring-necked parakeets and potential nest site competitors on surveyed sites and was therefore comparable with Strubbe & Matthysen (2007). For this a repeated measures generalized linear model with Poisson errors and log link was used, applied using the GENMOD procedure in SAS (SAS Institute 2001) to look at the relationship between ring-necked parakeets and each potential competitor in turn. Non-independence of successive counts in the same 1- km squares was taken into account by applying a repeated statement using 1-km square as subject.

It is important to control for species-habitat associations in such models as this may result in significant correlations unrelated to species interactions. All analyses were therefore repeated, controlling for the proportion of each 1-km survey site occupied by human habitation as obtained from the Centre for Ecology and Hydrology (CEH) Landcover dataset (Haines-Young et al . 2000).

A second set of analyses, investigated the possibility that populations of native cavity nesting birds have been depressed by increases in the abundance of ring-necked parakeets. Here a recent approach developed by Freeman & Newson (2008) was used, which estimates change in the abundance of a cavity nesting species in relation to the abundance of its potential competitors at the individual site level. Here change in abundance of each native species in relation to the number of parakeets was considered, controlling for potential competitive affects of other cavity nesting species.

4.2 Additional survey data analysis While none are as long-running as the BBS dataset, a number of additional surveys collect information about relative numbers of parakeets and native cavity nesting species during the breeding season and the following datasets were also analysed:

• The London Bird Project (2002-2004), a BTO project funded by the Bridge House Estate Trust, collected data on the abundance of bird species in summer and winter at 300 green space sites (including parks, cemeteries, woodland, marshland and common land) across the capital using a point-count technique. Of these, 70 sites were surveyed during the period April-June and ring-necked parakeets were recorded on eight of them. A generalized linear model with Poisson errors and log link, applied using the GENMOD procedure in SAS (SAS Institute 2001), was used to look at the relationship between maximum counts of ring-necked parakeets and those of each potential competitor in turn across these 70 sites.

• BTO/RSPB/BWI/SOC BirdTrack (2004-09) allows users to define their own sites and record numbers of birds seen at each on any day of the year. Sampling methods are not therefore standardised but is compensated for by

SID 5 (Rev. 07/10) Page 7 of 37 the relatively large volume of information. Data from 110 1-km squares from which ring-necked parakeets were recorded at least once were included in the analysis. As many sites were surveyed in successive years, a repeated measures generalized linear model with Poisson errors and log link was used, applied using the GENMOD procedure in SAS to look at the relationship between maximum counts of ring-necked parakeets and those of each potential competitor in turn. Non-independence of successive counts in the same 1-km squares was taken into account by applying a repeated statement using 1-km square as subject.

• BTO Garden BirdWatch (2006-09) allows users to record the maximum weekly count of each bird species seen in their garden throughout the year. Data for 276 gardens in which ring-necked parakeets were recorded at least once during the study period were included in the analysis. As many sites were surveyed in successive years, a repeated measures generalized linear model with Poisson errors and log link was used, applied using the GENMOD procedure in SAS to look at the relationship between maximum counts of ring-necked parakeets and those of each potential competitor in turn. Non-independence of successive counts in the same 1-km squares was taken into account by applying a repeated statement using 1-km square as subject.

As with the BBS analyses, the proportion of each monitored site that could be assigned to ‘human habitation’ on the basis of information from the CEH Landcover dataset (Haines-Young et al . 2000) was included in the models.

4.3. Results The first analysis explored the relationship between the number of ring-necked parakeets and the number of potential nest site competitors on surveyed sites. This approach is broadly comparable with that used by Strubbe & Matthysen (2007), who identified a negative correlation between ring-necked parakeet and nuthatch abundance across 44 sites in their Belgian study area.

In agreement with the findings of Strubbe & Matthysen (2007), the initial analyses of BBS data identified a significant negative association between numbers of ring-necked parakeets and nuthatches (P = 0.040, Table 4.1). However, when a variable specifying the proportion of the site covered by human habitation was included in this model, the relationship between the abundance of these two species was no longer significant. This suggests that the correlation identified is due to urban areas being positively associated with numbers of parakeets and negatively associated with nuthatches. Relationships between parakeet abundance and that of all other native cavity nesters were either non- significant or significantly positive.

Table 4.1. Cavity nesting species, information relevant to the analyses of BBS data and results of a simple comparison of ring-necked parakeets and focal species counts on BBS squares with and without controlling for habitat (% woodland and human habitation). P-values are: *<0.05, **<0.01, ***<0.001. Significant results are further highlighted in bold. Columns are estimated coefficients for the predictors, and their standard errors in parentheses.

Relationship with ring-necked parakeet counts No. of Mean count / Witho ut Controlling for Focal species sites in site (range) Trend 1 controlling for human habitat analyses human habitat Blue tit, Cyanistes caeruleus (BT) 171 12.4 (0 - 98) -1 1.291 (0.004)*** 0.003 (0.001) Green woodpecker, Picus viridis (G.) 145 1.1 (0 - 11) 117 ↑ 0.007 (0.003)** 0.009 (0.004)* Great spotted woodpecker, Dendropcopos major (GS) 155 1.1 (0 - 20) 97 ↑ 0.003 (0.005) 0.003 (0.006) Great tit, Parus major (GT) 172 6.4 (0 - 44) 25 ↑ 0.012 (0.004)** 0.007 (0.001)*** Jackdaw, Corvus monedula (JD) 110 4.5 (0 - 136) 200 ↑ 0.012 (0.006)* 0.011 (0.005) Nuthatch, Sitta europaea (NH) 83 6.4 (0 - 9) 17 -0.078 (0.040)* 0.006 (0.011) Ring-necked parakeet, Psittacula krameri (RI) 180 2.6 (0 - 226) 534 ↑ NA NA Stock dove, Columba oenas (SD) 88 0.7 (0 - 42) 31 -0.008 (0.013) 0.007 (0.006) Starling, Sturnus vulgaris (SG) 171 33.0 (0 - 346) -42 ↓ 0.031 (0.002)*** 0.001 (0.003)

An equivalent analysis was also carried out using the three additional datasets specified in Section 4.2 (Table 4.2). The only significant negative relationship identified between parakeet numbers and those of native cavity nesting species was for great tit when analysing the BirdTrack dataset, and even this did not persist once the proportion of human habitat had been accounted for in the model.

A second set of analyses used a recent approach developed by Freeman & Newson (2008) to evaluate the BBS dataset for evidence for rates of change in native cavity-nesting populations being depressed by parakeets (Table 4.3). Here, the only significant negative relationship identified between parakeets and a potential competitor was for blue tits. However, a coefficient value for α of -0.001 corresponds to an approximate 0.1% decrease in growth rate of blue tits on a BBS square for each additional parakeet, which is negligible and unlikely to shift this species from a positive to negative population growth.

SID 5 (Rev. 07/10) Page 8 of 37 Table 4.2. Results of a simple comparison of ring-necked parakeets and focal species counts at London Bird Project, BirdTrack and Garden BirdWatch sites with and without controlling for habitat (% human habitation). P-values are: *<0.05, **<0.01, ***<0.001. Significant results are further highlighted in bold. Columns are estimated coefficients for the predictors, and their standard errors in parentheses.

Relationship with ring-necked parakeet counts Focal species Without controlling Controlling for human for human habitat habitat LONDON BIRD PROJECT (N = 80) Blue tit 0.881 (0.111) *** 0.431 (0.119) ** Green woodpecker 0.056 (0.231) 0.268 (0.230) Great spotted woodpecker -0.210 (0.289) 0.047 (0.285) Great tit 0.238 (0.197) -0.112 (0.209) Jackdaw 0.198 (0.205) 0.778 (0.207) ** Nuthatch -1.220 (0.630) -0.494 (0.579) Stock dove 0.107 (0.221) -0.487 (0.579) * Starling 0.154 (0.212) -0.251 (0.227) BTO/RSPB/BWI/SOC BIRDTRACK (N = 110) Blue tit 0.010 (0.001) *** 0.005 (0.001) *** Green woodpecker 0.004 (0.002) * 0.011 (0.002) *** Great spotted woodpecker -0.005 (0.005) 0.005 (0.004) Great tit -0.065 (0.024) ** -0.027 (0.017) Jackdaw 0.004 (0.002) * 0.001 (0.001) Nuthatch 0.004 (0.002) * 0.004 (0.002) * Stock dove -0.004 (0.005) 0.001 (0.004) Starling 0.016 (0.000) *** 0.010 (0.000) *** GARDEN BIRDWATCH (N = 276) Blue tit 0.005 (0.011) 0.012 (0.010) Green woodpecker 0.033 (0.018) * 0.045 (0.019) ** Great spotted woodpecker 0.020 (0.009) * 0.030 (0.009) ** Great tit 0.006 (0.006) 0.010 (0.006) Jackdaw -0.017 (0.025) 0.009 (0.026) Nuthatch 0.040 (0.016) 0.0 58 (0.020) * Stock dove 0.043 (0.023) 0.048 (0.023) Starling 0.019 (0.009) * 0.020 (0.009) *

Table 4.3 Change in abundance of cavity-nesting species in relation to parakeet abundance according to Breeding Bird Survey data 1994-2008. P-values are: *<0.05, **<0.01, ***<0.001. Significant results are further highlighted in bold and results where the LR test is not significant are highlighted in italics. Columns are estimated coefficients for the predictors, and their standard errors in parentheses.

Change in population Native species growth rate Blue tit -0.001 (0.000)* Green woodpecker -0.002 (0.001) Great spotted woodpecker 0.000 (0.001) Great tit -0.001 (0.001) Jackdaw -0.001 (0.001) Nuthatch -0.002 (0.001) Stock dove 0.000 (0.003) Starling -0.001 (0.001)

4.4 Discussion Analyses of extensive national bird monitoring data from England provided no statistical evidence that English populations of nuthatch or other cavity-nesters are smaller in areas of higher parakeet density, although there is some evidence for a marginal depression of blue tit population growth rates with increased parakeet abundance. These findings are in contrast to those of Strubbe & Matthysen (2007), who identified a significant negative association between parakeet and nuthatch numbers across a number of study sites in Brussels and interpreted this as most likely to represent a causal relationship of one species on the other. While their models did not incorporate the proportion of each site that could be classified as ‘urban’, they did account for broad scale habitat type by including a term specifying the amount of woodland, which is likely to display a strong negative correlation with urban land cover. Strubbe & Matthysen (2009) subsequently undertook a cavity-exclusion experiment which showed a decrease in the number of nuthatch breeding pairs following blocking of parakeet nesting cavities. This was interpreted as evidence that supported the role of competition for nest-sites and explained the negative association between nuthatch and parakeet abundance. The results of this latter study, however, are not unequivocal as experimental and control sites differed in habitat, sample sizes of nuthatch pairs were small and numbers of parakeet pairs also decreased in one of the two treatment sites.

Assuming that the parakeets do competitively exclude nuthatches in Belgium, leading to a reduction in abundance, why are similar impacts not observed in England? It is possible that the ability to detect such an effect may have been

SID 5 (Rev. 07/10) Page 9 of 37 confounded by spatial variation in environmental variables across the study area. However, this is unlikely as several positive relationships between parakeet numbers and those of other cavity nesters were detected. The statistical power to detect a relationship using the Freeman & Newson (2008) model is relatively high, as indicated by the small, but significant, effects reported in Annex 1. Also the possibility cannot be excluded that the intensity of competition in England is reduced because the number of competing individuals is lower. A similar discrepancy might be predicted if the number of suitable cavities per unit area was lower in Belgium than in England (see sections 7.3 and 7.4 comparing cavity densities).

In conclusion, these analyses provided no strong evidence that, within their current geographical range and densities in England, ring-necked parakeets are negatively influencing the abundance of nuthatches or any other common native cavity nesting species. However, while no evidence was found of a population level effect, the possibility cannot be excluded that this threshold could have been exceeded at a subset of sites in England. Models used by Strubbe & Matthysen (2007) controlled for tree species composition and cavity availability that were identified as important in determining nuthatch abundance. Equivalent data are not currently available for the sites used in our analyses, although it should be noted that the deviance (Pearson Chi-square divided by the degree of freedom) of a model exploring the relationship between nuthatch numbers and parakeet abundance was close to one, thus providing little evidence for over-dispersion in the analyses, suggesting that the amount of unexplained variance in the dependent variable was low.

To investigate the possibility that parakeets may have a localised impact, a series of surveys was undertaken at 20 parkland sites in the London area at which detailed habitat data were also collected, as detailed in Sections 5 & 6. This also presented the opportunity to sample during the core parakeet breeding season, which starts and finishes much earlier than the BBS fieldwork period, as the latter is designed to cover the period of the year at which the majority of species, including late-arriving migrants, are breeding.

5. Pilot fieldwork

5.1 Study site selection The program MAXENT, which uses a maximum entropy modelling approach (Phillips et al. 2006), was used to predict parakeet presence from BTO/RSPB/BWI/SOC BirdTrack dataset (presence-only data) and habitat data from the LandCover Map 2000 dataset. When the same approach was applied to nuthatch, it was apparent that the limiting factor in determining whether parakeets had a negative impact on the abundance of this species was in finding a sufficient number of sites in the parakeet home range where nuthatches were also present. Many areas in which parakeets were predicted to occur at high densities were also difficult to survey accurately due to problems of access (e.g. suburban/urban areas).

To maximise the probability of identifying a significant interaction between these species should it exist, it was therefore necessary to focus on those sites in which habitat was potentially suitable for both, e.g. areas of parkland with low densities of mature trees. Restricting sites to a single habitat type also helped to control for potential confounding effects of this variable on species abundance relationships. As BTO Common Bird Census (CBC) data suggested that nuthatch densities in optimal habitat (broad-leaved woodland) are approximately one individual per 0.1 km 2, sites of less than 2.0 km 2 were excluded. Twenty-four sites in the London area (defined as within a 50km radius of the city centre) were selected on the basis of habitat type, with a mean area of 4.2 km 2 (range 1-18km 2).

Time constraints and access restrictions meant that not all sites could be surveyed as part of the pilot phase of the project, but the majority were visited to confirm habitat suitability, with the exception of a few at which parakeet presence had been confirmed using existing sources of information (e.g. Richmond Park, Windsor Great Park, Kew Gardens). Of the 24 original sites, two were excluded because they were too heavily wooded, therefore unsuitable for parakeets, and another two were ruled out because access permission would be impossible to obtain. Pilot survey data were collected at 12 of the remaining sites.

5.2 Field testing of techniques

5.2.1 Bird surveys Pilot surveys were undertaken to compare three different survey techniques at each site:

(i) Territory mapping – CBC methodology was used to record the locations of any individuals detected onto a site map. Registrations collected over a series of visits allowed individual territories to be identified. (ii) Line transects – transects ran parallel to the longest axis of the study site at intervals of 500m and all individuals detected were recorded in distance bands as per BTO/JNCC/RSPB BBS methodology (Risely et al . 2009). (iii) Point counts – at 500m intervals along each transect, following a two minute ‘settling period’, all birds seen or heard during a five-minute period were recorded, as recommended by Buckland (2006), using the same distance band technique as was used for the line transect methodology.

SID 5 (Rev. 07/10) Page 10 of 37 In addition to ring-necked parakeet, surveyors also recorded the six native cavity nesters included in Strubbe & Matthysen (2007): great spotted woodpecker, green woodpecker, jackdaw, nuthatch, starling and stock dove. At a subset of sites, blue tit and great tit were also recorded. In the subsequent intensive field surveys (section 6), all of these species plus lesser spotted woodpecker and coal tit Periparus ater (CT) were recorded. Again, however, lesser spotted woodpecker was too scarce for inclusion in analyses of abundance.

Territory mapping proved problematic and was only trialled at a few of the initial sites. The flocking nature of many species, particularly ring-necked parakeets, jackdaws and starlings, meant that the majority of individuals seen were observed in large groups in flight and could not be associated with individual territories. Results from line transect and point count techniques are presented in Table 5.1. While the point counts undertaken in the present study were of the same duration (five minutes) as those carried out during the Belgian study (Strubbe & Matthysen 2007), it is difficult to compare them directly as the results reported in the Belgian paper are limited to a subset of sites in ‘suitable habitat’ and insufficient information is given about the classification of habitat suitability. All surveyed sites have been included in Table 5.1, which may explain the relatively lower average numbers of individuals detected during point counts in the UK across all species except jackdaw.

Table 5.1. Mean numbers of each species recorded per 500m transect section using the two survey methods and proportion of sites and transects (trans) on which each species was recorded.

% % Line transect Point count

sites trans N Mean ± SE N Mean ± SE Blue tit 100 100 15 1.8 ± 0.4 13 0.3 ± 0.1 Great spotted woodpecker 83 64 22 0.2 ± 0.1 20 0.1 ± 0.1 Great tit 100 93 15 0.9 ± 0.3 13 0.4 ± 0.1 Green woodpecker 75 59 22 0.2 ± 0.1 20 0.2 ± 0.0 Jackdaw 100 91 22 9.3 ± 1.9 20 3.0 ± 0.7 Nuthatch 25 23 22 0.1 ± 0.1 20 0.0 ± 0.0 Ring-necked parakeet 42 36 22 1.4 ± 0.5 20 0.4 ± 0.2 Starling 50 54 22 1.5 ± 0.6 20 0.4 ± 0.2 Stock dove 58 50 22 0.2 ± 0.1 20 0.1 ± 0.0

Overall, the transect methodology gave a higher detection rate than that achieved using point counts, permitting a more powerful analysis, but unlike the territory mapping approach, it does not provide information on potential nest site location for the more intensive fieldwork phases of this project. The approach adopted for the year two (2010) breeding season, therefore, was a hybrid method using line transects to monitor abundance while also noting the position of target species observed on site maps. As the transect methodology was less time-consuming than territory mapping, this approach allowed fieldworkers to undertake targeted nest searches.

5.2.2 Power analysis The effect size and standard errors of the relationship between ring-necked parakeet and nuthatch numbers reported by Strubbe & Matthysen (2007) in Belgium were obtained from the authors. The data collected during the present pilot surveys were then used to calculate the probability of identifying an effect of similar magnitude given the levels of abundance of these species detected at our study sites. The power analysis was carried out using a repeated measures framework, where the dependent variable was the number of nuthatches observed on each sampling occasion and the independent variable was the number of parakeets observed. The number of survey visits and sites surveyed varied between models, with 2000 simulations run for each combination.

The results indicated that increasing the number of survey 4 visits visits from two to four per annum increased the probability of 3 visits detecting an effect by 15-20%, while increasing the number of 2 visits sites by 10 increased the power by c.8% (Figure 5.1). Further modelling showed than an increase in the number of sites surveyed beyond 28 had little additional effect on the power statistical power of the analysis. The statistical power is limited by the small size of the effect identified in the Belgian study and the relatively low density of nuthatches within the core range of the parakeets. 0.0 0.2 0.4 0.6 0.8 1.0 12 14 16 18 20 22 24 Based on the results of the power analysis, the decision was number of sites made to select a minimum of 20 study sites at which four Figure 5.1 . Power to detect an effect of ring- survey visits would take place across the breeding season; necked parakeet on nuthatches of similar this was subsequently reduced to three visits per site, the last magnitude to that observed in the Belgian study visit in June being dropped to avoid inflation of apparent increase with the number of sites and the parakeet breeding numbers by the inclusion of fledged number of visits to each site. juveniles in counts. The majority of these sites were those

SID 5 (Rev. 07/10) Page 11 of 37 selected during the pilot surveys, but some had to be dropped because of refusal of landowners to grant access or habitat unsuitability. Additional sites were selected that were all within the same overall study area (50km radius of the centre of London) and were of the same habitat type (areas of parkland with low densities of deciduous trees).

6. Species abundance

6.1 Methodology Over the winter of 2009/2010 access permission was arranged for all 20 study sites (see Appendix I for locations) and transect routes established. Transects ran parallel to the longest axis of the study site at intervals of 500m. The numbers of transects and overall transect length varied between sites dependent on the areas of the individual sites (range = 500-5700m). During a survey all individuals seen or heard were recorded in distance bands on either side of the transect line (within 25m, between 25m and 100m, beyond 100m; birds in flight were also recorded at any distance) as per BBS methodology (Risely et al . 2009). The first round of bird surveys was initiated in mid-Feb 2010 and lasted for approximately 2-3 weeks. Subsequent bird surveys were conducted during mid-March and mid-April. Parakeet abundance was expressed as the maximum count during any one of the survey visits (across all distance bands, including birds in flight) and also as the median count.

Habitat characteristics were recorded using a combination of field surveys and desk-based GIS analysis. Visits were made to each site at the end of the breeding season to record the number of veteran trees, defined as having a circumference >5m. ArcMap 10 (ESRI, Redlands, California, US) was used to import aerial photographic images of each site from GoogleMaps. The area of each site was calculated, as was the area within the site boundary covered by each of six main habitat types: (i) continuous closed canopy woodland- no gaps in canopy, (ii) medium density woodland – gaps in canopy, but area of trees >area of open ground, (iii) low density woodland – gaps in canopy, but area of trees

Analysis investigated any effect of parakeet abundance on the abundance of native species, as performed for the existing BTO datasets outlined in Section 4.3. A generalised linear model (GLM, Poisson distribution, log link function) was fitted, with the median number of observations of native species as the dependent variable and the median number of parakeets as the explanatory variable. All models were then re-run including terms for the three habitat characteristics thought to have greatest potential to influence cavity availability: the proportion of ‘continuous closed canopy woodland’, ‘medium density woodland’ and the ‘density of veteran trees’. It was not possible to include further habitat variables in the model due to limitations of sample size. Due to this and collinearity between some habitat variables inflating some parameter standard errors, the significance of interaction terms between habitat and parakeet numbers were therefore modelled separately for each of the three selected habitat parameters.

6.2 Results

Species counts Numbers of ring-necked parakeets varied markedly between sites (maximum = 0-324, median = 0-320) (Table 6.1). Counts of native species also varied markedly between sites (e.g. nuthatch: maximum = 0-21, median = 0-16).

Table 6.1. Maximum bird counts recorded at study sites and total transect length.

Study Site RI NH SD G. GS LS BT GT CT SG JD Transect (m) Bedford Park 0 1 8 11 8 0 68 44 2 4 48 1400 Lamberhurst 0 9 18 7 11 0 106 46 3 4 93 3200 North Mymms 0 8 13 16 9 0 101 76 5 10 181 2600 Weald Country Park 0 6 9 9 7 0 61 39 5 10 179 1600 Cassiobury Park 3 6 20 13 20 1 117 66 7 19 35 2500 Silwood Park 6 8 7 9 11 0 62 19 8 6 53 900 Cliveden Estate 10 21 15 12 11 0 131 38 3 0 70 2600 Grovelands Park 13 5 13 4 10 0 51 26 0 57 1 900 Greenwich Park 18 4 5 1 7 0 49 13 4 33 7 700 Regents Park 18 0 3 3 6 0 63 20 0 105 0 1800 Knole Park 35 18 50 20 14 0 138 32 6 26 398 4200 Wimbledon Common 49 4 3 15 39 0 266 100 8 53 75 5700 Hampstead Heath 65 18 15 8 16 0 222 68 13 51 25 3700 Langley Park 73 3 7 15 16 0 125 43 5 75 343 2700 Windsor Great Park 76 17 23 19 13 1 100 29 6 40 1030 3900 Kew Gardens 84 1 12 12 8 0 97 38 5 65 22 2400 Central London Parks 85 1 13 4 4 0 81 31 4 133 0 2600 Bushy Park 216 14 13 31 15 1 119 79 3 176 417 4700 Osterley Park 227 6 6 10 6 3 61 32 7 51 70 1400 Richmond Park 324 17 28 32 25 1 170 54 0 136 1073 5600

SID 5 (Rev. 07/10) Page 12 of 37 Study sites varied in their total area and, therefore, in the overall length of transect (700-5,700m) along which birds were counted. Counts were standardised between sites by expressing parakeet abundance as encounter rates, i.e. parakeets per metre of transect. Figure 6.1 presents encounter rates based on median site counts.

Based on median counts, ring-necked parakeet abundance ranged from 0 to 0.057 parakeets per metre of transect. Overall, sites exhibited a continuous gradation in parakeet abundance. At the two extremes of the range, however, were sites with zero or very low (<0.01 parakeets per m) parakeet abundance and sites with the highest abundance (>0.025 parakeets per m). With reference to figure 2, Bedford to Wimbledon represented sites with zero/low abundance and Greenwich to Richmond represented the sites with the highest abundance of parakeets. Figure 6.1. Encounter rates for ring-necked parakeet across twenty London Comparison of nest site characteristics parkland study sites. Encounter rates based on the median number of birds and breeding parameters of native recorded across three transect surveys. species are compared between zero/low and high parakeet abundance sites in sections 7 and 8.

Restricting sites to one broad habitat type of ‘parkland’ helped control for potential confounding effects of broad habitat on species abundance relationships. However, there were differences between sites in the finer-scale habitat variables considered. Woodland variables, used in analyses of abundance relationships, are presented in Table 6.2.

Table 6.2. Summary of woodland variables for the study sites.

Woodland Total Continuous canopy Medium density Veteran Study Site Area (ha) ha % ha % trees Bedford Park 163.5 46.3 28.3 14.4 8.8 22 Lamberhurst 360.6 117.1 32.5 23.2 6.4 109 North Mymms 229.8 72.3 31.4 0.0 0.0 95 Weald Country Park 183.1 89.5 48.9 12.8 7.0 167 Cassiobury Park 103.4 15.8 15.3 39.9 38.6 92 Knole Park 567.1 14.2 2.5 229.5 40.5 105 Cliveden Estate 156.8 103.8 66.2 8.4 5.3 32 Regents Park 197.1 0.00 0.0 94.3 47.9 13 Silwood Park 140.7 59.6 42.4 14.7 10.5 56 Wimbledon Common 653.0 324.9 49.8 58.7 9.0 31 Grovelands Park 49.3 20.2 41.0 7.1 14.4 45 Langley Park 216.2 48.6 22.5 24.3 11.2 179 Hampstead Heath 453.0 154.5 34.1 46.7 10.3 97 Windsor Great Park 825.7 204.5 24.8 127.9 15.5 387 Greenwich Park 79.3 0.0 0.0 44.5 56.1 111 Central London Parks 290.0 0.0 0.0 118.9 41.0 91 Kew Gardens 286.0 51.5 18.0 58.3 20.4 58 Osterley Park 144.6 22.3 15.4 23.9 16.5 63 Bushy Park 762.7 35.4 4.6 115.5 15.1 159 Richmond Park 1095.9 221.5 20.2 182.8 16.7 369

Relative species abundance Analyses of the abundance relationships between native species and parakeets, without accounting for variation in habitat variables, revealed only one significant relationship - great tit numbers overall were positively correlated with parakeet numbers (Table 6.3).

Table 6.3 Summary of results of GLM investigating the relationships between numbers of parakeets and numbers of native species, without accounting for habitat variables. Significant results are highlighted in bold.

NH SD G. GS BT GT CT SG JD Df 1,18 1,18 1,18 1,18 1,18 1,18 1,18 1,18 1,18 Deviance ratio 0.60 0.05 3.70 0.65 2.74 5.45 0.01 0.94 0.74 F pr 0.45 0.83 0.07 0.43 0.12 0.03 0.92 0.34 0.40

SID 5 (Rev. 07/10) Page 13 of 37 Similar analysis but accounting for habitat variables ( ‘continuous closed canopy woodland’, ‘medium density woodland’ and the ‘density of veteran trees’) again failed to identify any significant negative relationships between the numbers of parakeets and native species (Table 6.4). As above, however, there was a positive relationship between the numbers of parakeets and the numbers of great tits.

Table 6.4 Summary of results of GLM investigating relationships between numbers of parakeets and numbers of native species, accounting for habitat - ‘continuous closed canopy woodland’, ‘medium density woodland’ and the ‘density of veteran trees’. Significant results are highlighted in bold.

NH SD G. GS BT GT CT SG JD Df 1,15 1,15 1,15 1,15 1,15 1,15 1,15 1,15 1,15 Deviance ratio 0.34 0.02 1.82 0.01 1.45 6.27 0.43 1.18 0.49 F pr 0.57 0.90 0.20 0.91 0.25 0.02 0.52 0.30 0.49

When examining interactions between parakeet numbers and separate individual habitat variables, however, for two species – green woodpecker and stock dove, there was evidence that a negative relationship with increasing parakeet abundance was significantly affected by ‘continuous closed canopy woodland’ (Table 6.5). Numbers of green woodpecker and stock dove decreased with increased numbers of parakeets, and the positive interaction between parakeet numbers and habitat indicated that this relationship was more pronounced with decreasing proportion of continuous closed canopy woodland.

Table 6.5 Summary of results of GLM analysis investigating changes (interaction) in the relationship between the numbers of native species and the numbers of parakeets for varying levels of i. ‘continuous closed canopy woodland’, ii. ‘medium density woodland’ and iii. ‘density of veteran trees’. P-values are: *<0.05, **<0.01, ***<0.001. Significant results are further highlighted in bold. Columns are estimated coefficients for the predictors, and their standard errors in parentheses.

Species ‘continuous closed ‘medium density ‘density of canopy woodland’ woodland’ veteran trees’ Nuthatch 0.0000444 (0.0000238) 0.000030 (0.000199) 0.00000365 (0.000040) Stock dove 0.0000324 (0.0000151) * 0.000133 (0.000235) -0.00000126 (0.00000318) Green woodpecker 0.0000381 (0.0000170) * 0.000006 (0.000145) -0.00000076 (0.00000252) Great-spotted woodpecker - a 0.000159 (0.000105) -0.00000241 (0.00000205) Blue tit - a 0.000143 (0.0000794) -0.00000332 (0.00000141) Great tit - a 0.0001076 (0.0000725) -0.00000213 (0.00000161) Coal tit - a -0.0001810 (0.0000801) -0.00000221 (0.00000303) Starling -0.000014 (0.0000264) 0.000227 (0.000323) -0.00000759 (0.00000380) Jackdaw 0.0000132 (0.0000311) 0.000251 (0.000432) 0.0125 (0.00000350) a Due to collinearity, when fitting interactions the parameter standard errors were severely inflated, therefore it was not possible to differentiate the effect of the two main parameters from that of the interaction.

6.4 Discussion Analyses of abundance data from 20 parkland sites in the Greater London area indicated no overall negative relationships between parakeet numbers and numbers of native species. Further analyses that took into account variation in selected habitat variables (considered most likely to influence cavity availability) between sites confirmed no negative relationships between the numbers of parakeets and the numbers of native species. There was, however, a positive relationship between parakeet numbers and numbers of great tits, i.e. numbers of great tits increased with increased numbers of parakeets.

Due to small sample sizes, the interactions between parakeet numbers and habitat variables, thought to have greatest potential to influence cavity availability, were investigated individually. A decrease in the numbers of green woodpecker and stock dove with increased numbers of parakeets was more pronounced with smaller proportion of continuous closed canopy woodland. There was no similar relationship for either of the two other habitat variables (‘medium density woodland’ and ‘density of veteran trees’ ). This relationship is consistent with some degree of competition between parakeets and green woodpecker and stock dove in sites with a low proportion of closed canopy woodland. Less wooded habitat might be predicted to provide fewer cavities, resulting in increased competition for nest sites, but the possibility cannot be excluded that such habitat also holds relatively greater numbers of green woodpeckers and stock doves, increasing the statistical power to detect a significant impact of parakeets.

Therefore, although there was no decrease in numbers of green woodpeckers and stock doves at the national scale associated with parakeet abundance, there may be a local scale effect on these species. For all other species studied, there was no strong evidence for a significant detrimental effect of parakeet numbers on the numbers of native cavity- nesters at either national or local-scale.

As for the regional scale analysis (section 4), explanations for the general absence of a negative impact are that the intensity of competition in England is likely to be reduced if the number of competing individuals is lower and/or the number of suitable cavities per unit area is higher in England, compared to Belgium where a negative effect of parakeet abundance on nuthatch numbers was observed. Data presented in Section 7 shows that cavity abundance is higher in the present study sites compared to those in Belgium.

SID 5 (Rev. 07/10) Page 14 of 37 Taken together, these analyses may indicate that the key nesting resource, cavity abundance, is high enough to support both parakeets and native species, without any overall significant detrimental effects of parakeet on native species. It is stressed, however, that the general absence of an effect of parakeet abundance on the numbers of native species is relevant to the parkland sites studied at the present time. In habitats where the relative abundance of parakeets, native species and nesting resources facilitate increased competition for nest sites, there remains the potential for detrimental impacts.

7. Nest habitat

7.1 Srubbe & Matthysen (2007) show that parakeets do have the potential to depress populations of cavity nesters but the mechanism is as yet unclear. There is much evidence to suggest that availability of nest cavities, a declining resource in many areas of the UK due to the loss of old buildings and trees, can limit the population size of secondary cavity nesters (Newton 1994, Strubbe & Matthysen 2007). The relatively large size and aggressive nature of parakeets suggest that they may be able to out-compete many native species for suitable nesting locations through direct usurpation of tree cavities. In addition, parakeets nest much earlier than native species, often commencing in January, and, due to the relatively long incubation and nestling periods (c. 2.5 months from laying to fledging, Cramp 1985), cavities may already be occupied when native species are prospecting for nesting sites and commencing breeding. In the latter case, competitive exclusion as a result of early parakeet breeding and associated early occupancy of preferred cavities can be predicted to have a greater potential impact if parakeets and native species share preferences for nest-cavity characteristics. In the absence of access to a suitable nest cavity, native species may respond by delaying or foregoing breeding. Conversely, if there is little overlap in shared preferences for nest sites, there will be less likelihood of an impact on nest site acquisition and breeding.

The following fieldwork investigated overlap between parakeets and native species in nest site characteristics. In addition, nest site characteristics were compared between groups of sites with the lowest and highest abundance of parakeets respectively. If parakeets are competitively excluding native species from preferred cavities, it may be predicted that the characteristics of native species nest sites will vary between the two groups of sites.

7.2 Methodology

Nest distribution Intensive searches of 15 of the 20 study sites were conducted regularly throughout the breeding season (February to July inclusive) to locate and map the nest sites of all target species.

Nest site habitat At each nest, site habitat characteristics were measured at three spatial levels: (i) the nest cavity, (ii) the nest tree, and (iii) the habitat immediately surrounding the tree (Table 7.1). For (iii) features were measured using a circular sample- plot method (James & Shugart 1970) - sample plots were of a radius of 12.62m (0.05ha) centred on the nest tree.

Table 7.1 Summary of nesting habitat characteristics.

Level Variable Unit Description Cavity Cavity origin W/N Woodpecker or Natural (e.g. loss of limb) Cavity position T/B Trunk or Branch Cavity height m height of cavity above ground (measured with a clinometer) Cavity orientation ºN compass bearing that entrance faces Minimum entrance diameter cm size category <4cm or 4-8cm or >8cm (estimated) Aspect A/O/V aspect to vertical (Acute, Obtuse, Vertical) Tree Species common ( latin ) tree species Status L/D tree Live or Dead Height m height of tree (measured with a clinometer) Diameter at breast height (DBH) m circumference of tree measured at 1.4m from the ground (D= C/ π) No. Cavities No. total number of cavities in nest tree (by size category) Canopy depth m max. minus min. canopy height (measured with a clinometer) Habitat Global crown cover % % cover of the crowns of all trees within the sample plot Global ground cover % % ground cover of all shrubs within the sample plot Mean tree diameter (DBH) m mean DBH of 10 trees (>10cm) nearest centre of plot (inc. nest tree) th 2 Tree density (D) trees/ha distance to 10 tree >10cm diameter (D=100,000 / ( π * dmax ) Cavity density (nest tree plot) No./ha total cavities (by size category) found during 5 min. search Cavity density (random tree plot 1) No./ha total cavities (by size category) found during 5 min. search 1Random plots were selected using a random compass bearing and distance from the nest tree.

Cavity abundance The impact of ring-necked parakeets on native hole-nesting species is likely to vary with the overall availability of suitable cavities. Counts of tree cavities were undertaken at nest tree and random non-nest tree plots in each of the 15 study sites. At each plot, following Strubbe & Matthysen (2009), a count was undertaken from the ground using

SID 5 (Rev. 07/10) Page 15 of 37 binoculars when required. Each cavity was allocated to one of three size categories according to estimated entrance diameter (<4cm, 4-8cm, >8cm). In the present study individual plots were searched for 5 minutes, in comparison to 10 minutes per plot in Strubbe & Matthysen’s (2009) study. Mean cavity density was estimated for each site, expressed as cavities per ha, calculated as the mean of the averaged cavity counts for pairs of nest tree and random non-nest tree plots (0.05ha per plot).

Statistical analyses Differences in nest site habitat variables were investigated using ANOVA for continuous variables (e.g. tree height) and GLM for binary variables (e.g. nest in the trunk vs. in the branches). Habitat variables were tested, in turn, for differences: (i) between species, and (ii) within individual species between sites with zero/low parakeet abundance and high parakeet abundance (as categorised in Section 6.3). For categorical variables, differences in the relative proportions of nests located in each of the alternate categories for each variable (e.g. cavity origin - cavity located in the trunk or in a branch) were investigated using two-sample binomial tests. For continuous variables, ANOVA followed by pairwise comparisons were used to identify individual differences between species. The analysis aimed to identify overlap in nesting characteristics of ring-necked parakeets and native species, and to investigate whether native species nesting habitat varied with parakeet abundance (i.e. between zero/low parakeet abundance sites and high parakeet abundance sites). Statistical analyses did not involve adjustment of significance levels for multiple tests. As the study aimed to identify any potentially detrimental effects of parakeets on native species, the risk of a Type I error (false positive) was considered to be of less concern than Type II error (false negative).

7.3 Results

Nest distribution A total of 235 nests of nine different species were discovered (Table 7.2). For a number of species the distribution of nests between different sites was markedly uneven. For example, 36% of all starling nests were discovered in one site – Cassiobury. There were 22 trees in which multiple nests were present. In six cases, ring-necked parakeets shared the same nest tree with a native species – jackdaw (2), starling (2), nuthatch (1) and green woodpecker (1).

Table 7.2 Numbers of nest cavities of different species recorded at the different study sites.

Site RNP G. GS JD NH SD SG GT BT Tot Bedfords 0 0 1 6 1 0 0 0 0 8 Lamberhurst 0 1 2 1 1 0 0 0 0 5 Weald 0 1 3 8 1 0 3 0 0 16 Cassiobury 0 0 1 3 2 1 12 0 1 20 Silwood 0 0 2 6 3 1 0 0 0 12 Knole 1 1 2 10 2 2 5 0 0 23 Cliveden 1 0 1 1 2 0 0 0 0 5 Wimbledon 5 1 7 0 2 1 0 1 0 17 Windsor 8 1 1 7 0 0 0 0 1 18 Langley 2 0 1 5 1 0 0 1 0 10 Kew 5 0 1 0 0 2 0 2 0 10 Hyde 6 2 0 0 2 2 7 0 0 19 Osterley 4 2 0 1 1 1 0 1 0 10 Bushy 17 1 3 8 0 0 0 2 0 31 Richmond 13 2 2 6 2 0 6 0 0 31 Total 62 12 27 62 20 10 33 7 2 235

Categorical variables Categorical variables analysed were: cavity origin (woodpecker or natural), cavity position (trunk or branch), entrance diameter (<8cm or >8cm), aspect (vertical or acute/obtuse) and tree status (live or dead). Sample sizes were only large enough to include four native species in the analyses: great-spotted woodpecker, jackdaw, nuthatch and starling. Nests of the target species (parakeets and native species) were located in trees of 19 genera, with oak Quercus sp. accounting for the most nests for both ring-necked parakeets and native species (Table 7.3).

Table 7.3 Summary of relative percentages of cavity origin (woodpecker or natural), cavity position (trunk or branch), cavity entrance diameter (<8cm or >8cm), cavity aspect (vertical or acute/obtuse) tree status (dead or live) and tree species used by the target species.

Cavity origin Cavity position Entrance di am. Cavity aspect Tree status Tree species Species % Woodpecker % Trunk %<8cm % Vertical % Dead % Quercus Ring-necked parakeet 77.4 45.2 75.8 59.7 6.5 77.4 Stock dove 20.0 50.0 30.0 80.0 20.0 60.0 Green woodpecker 100.0 83.3 100.0 75.0 16.7 58.3 Great-spotted woodpecker 100.0 66.7 100.0 74.1 33.3 37.0 Jackdaw 22.6 53.2 32.3 82.3 12.9 56.5 Great tit 71.4 71.4 85.7 71.4 14.3 42.9 Nuthatch 25.0 35.0 100.0 85.0 0.0 55.0 Starling 63.6 39.4 97.0 72.7 3.0 68.8

SID 5 (Rev. 07/10) Page 16 of 37 There was significant variation between bird species in three categorical variables – cavity origin, entrance diameter and tree status. In comparison to parakeets, all four native species exhibited overlap (i.e. no significant difference) in one or more categorical habitat variables (Table 7.4). Jackdaw showed the most differences in habitat variables to parakeets, showing more frequent use of a larger cavity entrance diameter, naturally formed (rather than woodpecker) cavities and cavities aligned at an angle rather than vertical. Nuthatch also showed significantly more frequent use of natural cavities than parakeets. All nuthatch cavities and zero parakeet cavities were estimated to be <4cm.

Table 7.4 Summary of comparison of categorical habitat variables between species (GLM with binomial distribution, logit- link function).

Comparison between species Comparison with RI Deviance Variable df ratio F pr GS JD NH SG Natural v Woodpecker cavity 4,47 18.47 <0.001 0.623 <0.001 <0.001 0.137 Branch v Trunk 4,47 0.78 0.546 0.315 0.565 0.506 0.592 Entrance diameter <8cm v >8cm 4,47 16.57 <0.001 0.662 <0.001 0.118 0.111 Vertical v Acute/Obtuse 4,47 1.58 0.196 0.305 0.034 0.115 0.318 Proportion in dead trees 4,47 6.88 <0.001 <0.001 0.156 0.744 0.405

Native species almost universally failed to exhibit differences in categorical habitat variables between sites with zero/low parakeet abundance and sites with high parakeet abundance (Table 7.5); the only differences being great- spotted woodpecker and starling having a higher proportion of nests in branches (rather than trunks) in sites with high parakeet abundance. The most frequent difference in habitat features were shown by parakeets themselves. In high abundance, parakeets more frequently utilised cavities that were in branches (rather than the trunk), with a smaller entrance diameter (<8cm), were aligned at an angle (rather than vertical) and in live (rather than dead) trees.

Table 7.5 Summary of comparison of categorical habitat variables for individual species between sites with high parakeet abundance and sites with zero/low parakeet abundance (two-sample binomial tests).

Species Variable RI (45,7) GS (6,19) JD (15/35) NH (5/14) SG (13/20) Branch v Trunk 0.008 0.037 0.304 0.363 0.023 Entrance diameter <8cm v >8cm 0.035 1a 0.280 0.539 0.143 Natural v Woodpecker cavity 0.553 1a 0.131 0.709 0.346 Vertical v Acute/Obtuse 0.031 0.169 0.849 0.372 0.245 Proportion in dead trees 0.026 0.258 0.849 1a 0.208 aProportions for parakeet and native species were either both 0 or both 1. Sample sizes in parentheses (high parakeet, zero/low parakeet)

Continuous variables Continuous variables analysed were: cavity height, tree height, tree diameter at breast height (DBH), canopy depth, crown cover, ground cover, mean DBH and tree density. The nests of ring-necked parakeets were located in trees with a mean height of 21.2 ± 5.97m (n=62), a mean DBH of 1.04 ± 0.40m (n=61) and a mean canopy depth of 17.8 ± 7.41m (n=62) (Table 7.6). The mean height of the nest cavity was 9.3 ± 3.28m (n=62).

Table 7.6 Summary of nest cavity, tree height, diameter at breast height, nest cavity height and canopy depth for nests of target species.

Cavity ht. Tree ht. DBH Canopy Crown Ground Mean Tree Species (m) (m) (m) depth (m) cover (%) cover (%) DBH (m) density Ring-necked parakeet 9.3 ± 3.28 21.2 ± 5.97 1.04 ± 0.40 17.8 ± 7.41 55.6 ± 2.87 25.4 ± 4.26 0.82 ± 0.04 75 ± 14.7 Stock dove 7.7 ± 4.05 19.7 ± 6.34 1.10 ± 0.51 16.3 ± 8.17 61.0 ± 4.82 41.0 ± 13.30 0.86 ± 0.16 128 ± 61.0 Green woodpecker 6.2 ± 2.80 18.5 ± 7.79 0.75 ± 0.25 12.3 ± 7.93 65.8 ± 5.60 44.6 ± 10.20 0.63 ± 0.11 180 ± 46.1 Great-spotted woodpecker 6.4 ± 2.71 15.6 ± 7.85 0.73 ± 0.39 10.4 ± 9.56 69.6 ± 4.27 43.8 ± 7.23 0.52 ± 0.06 341 ± 90.4 Jackdaw 7.2 ± 2.93 19.3 ± 6.30 1.16 ± 0.39 15.1 ± 8.05 61.6 ± 3.77 34.9 ± 4.62 0.83 ± 0.05 147 ± 19.9 Great tit 3.0 ± 0.71 14.3 ± 6.13 0.71 ± 0.31 8.4 ± 5.25 69.3 ± 12.60 49.3 ± 13.70 0.46 ± 0.07 160 ± 60.8 Nuthatch 10.2 ± 4.67 23.5 ± 4.82 0.94 ± 0.21 18.6 ± 5.82 72.5 ± 4.90 39.5 ± 9.21 0.68 ± 0.07 135 ± 25.0 Starling 8.9 ± 3.30 19.7 ± 4.90 1.21 ± 0.38 16.8 ± 6.04 55.6 ± 3.89 14.1 ± 4.24 1.11 ± 0.07 52 ± 81.3

In comparison to ring-necked parakeet, of the four native species only great-spotted woodpecker showed a significant difference in all continuous nest habitat variables (Table 7.7). The other three species (jackdaw, nuthatch and starling) exhibited no difference with parakeets in the majority of variables.

For all four native species, nests were associated with higher tree density in sites with high parakeet abundance compared to sites with zero/low parakeet abundance (Table 7.8). In addition, nuthatch occupied taller trees (also starling) with greater canopy depth in high abundance parakeet sites. Jackdaw occupied nests associated with lower crown cover in high abundance parakeet sites. Ring-necked parakeet nests were located in trees with greater diameter and lower crown cover in high abundance sites.

SID 5 (Rev. 07/10) Page 17 of 37 Table 7.7 Summary of pairwise comparisons (following ANOVA) examining differences in continuous variables between ring-necked parakeets and native species.

Significant difference to parakeet (p<0.05) Variable GS JD NH SG Nest height Yes ↓ Yes ↓ no no Tree height Yes ↓ no no no Tree diameter Yes ↓ no no Yes ↑ Canopy depth Yes ↓ Yes ↓ no no Crown cover Yes ↑ no Yes ↑ no Ground cover Yes ↑ no no no Mean tree diameter Yes ↓ no no Yes ↑ Tree density Yes ↑ no no no Arrows indicate the magnitude of the variable for the native species relative to ring-necked parakeet. e.g. for great-spotted woodpecker nest heights are lower than those of ring-necked parakeet.

Table 7.8 Summary of comparison of continuous habitat variables for individual species between sites with high parakeet abundance and sites with zero/low parakeet abundance (two-sample t-tests).

RI (45,7) GS (6,19) JD (15,35) NH (5,14) SG (13,20) Variable t p t p t p t p t p Nest height -0.87 0.388 -1.39 0.178 1.78 0.082 -0.17 0.868 0.29 0.775 Tree height -1.86 0.068 -1.24 0.227 -0.57 0.573 -2.18 0.044 ↑ -2.42 0.021 ↑ Tree diameter -3.47 0.001 ↑ -1.27 0.217 0.72 0.478 0.18 0.859 0.79 0.438 Canopy depth -1.71 0.093 -1.33 0.197 -0.80 0.428 -2.63 0.018 ↑ -1.03 0.318 Crown cover 2.17 0.035 ↓ 1.60 0.124 3.11 0.003 ↓ 1.57 0.134 1.87 0.071 Ground cover 0.83 0.413 -0.17 0.865 -0.08 0.933 0.61 0.549 -0.28 0.782 Mean tree diameter -3.42 0.001 ↑ -1.16 0.258 -1.51 0.138 -1.92 0.073 -0.41 0.687 Tree density 2.39 0.054 2.82 0.010 ↑ 4.98 <0.001 ↑ 3.89 0.002 ↑ 2.07 0.048 ↑ Sample sizes in parentheses (high, zero/low) except JD 35,13 tree diameter and mean tree diameter; SG 20,12 tree density. Arrows indicate the magnitude of the variable for high parakeet sites relative to zero/low parakeet sites; e.g. for nuthatch tree height was higher in high parakeet sites compared to zero/low parakeet sites.

The orientation of ring-necked parakeet cavities (n=62) did not exhibit a preference for any direction relative to compass bearing, i.e. the bearing of cavity entrances was uniformly distributed with compass bearing (Rayleigh’s Test for Circular Uniformity: Z 0.5,62 = 1.1771, 0>0.20).

Cavity abundance Cavity abundance in each of the study sites is presented in Table 7.9, with cavities categorised into three sizes according to estimated entrance diameter: <4cm, 4-8cm and >8cm. The median density of cavities with >4cm diameter entrance was 55 per ha (range: 22 – 94 per ha) – only nuthatch used cavities <4cm diameter.

Table 7.9 Mean estimated cavity density for the study sites.

Mean densities of cavities (no./ha) Median density of cavities >4cm (no./ha) na Site All Nest Random <4cm >4<8cm >8cm plots plots plots Bedford Park 6 8.3 33.3 15.0 48.3 96.7 0.0 Lamberhurst 5 6.0 48.0 20.0 68.0 132.0 4.0 Weald Country Park 14 8.6 30.0 29.3 59.3 88.6 30.0 Cassiobury Park 17 11.8 55.9 35.3 91.2 158.8 23.5 Silwood Park 9 10.0 35.6 17.8 53.4 91.1 15.6 Clivedon Gardens 5 12.0 22.0 0.0 22.0 36.0 8.0 Knowle Park 21 22.9 48.1 45.7 93.8 152.4 35.2 Wimbledon Common 17 8.8 23.5 11.2 34.7 64.7 4.7 Langley Park 8 3.8 13.8 50.00 63.8 95.0 28.9 Windsor Great Park 16 1.9 18.1 36.9 55.0 90.0 20.0 Kew Gardens 10 4.0 26.0 17.0 43.0 80.0 6.0 Hyde park 15 5.3 20.7 18.0 38.7 72.0 5.3 Bushy Park 19 6.5 32.5 24.0 56.5 91.0 37.7 Osterley Park 10 8.0 35.0 10.0 45.0 86.0 4.0 Richmond Park 25 6.8 41.6 30.0 71.6 99.2 44.0 Median 197 8.0 32.5 20.0 55.0 91.0 15.6 anest tree plot and random plot pairs

Nest tree plots held significantly higher numbers of cavities compared to paired random plots (nest tree plots: median = 91 per ha, random plots: median = 16 per ha, Wilcoxon Matched Pairs Test: Z=-10.628, n=183, p<0.001).

There was a positive but non-significant trend for the total abundance (encounter rate based on median counts) of cavity nesters (all species combined) to increase with higher cavity density (all cavity types) (Spearman Rank: Rs=0.458, n=15, p=0.09).

SID 5 (Rev. 07/10) Page 18 of 37 7.4 Discussion Relatively little information about nest-site selection in ring-necked parakeets is available in the literature. The mean nest height previously reported for the UK was 8.5 ± 4.1m (n=70), with the trees being a mean of 19.5 ± 8.1m tall (n=70) with an average DBH (tree diameter at breast height) of 73.7 ± 34.1cm (n=69; Butler 2003). The same study recorded parakeets breeding in trees of twelve different genera, with ash Fraxinus accounting for 33% of the trees used and oak accounting for 22%. Breeding attempts (n=108) were found principally in old woodpecker nests (53.7%), but were also found in natural cavities (28.7%) and nest boxes (17.6%).

In the present study, the broad characteristics of parakeet nest trees were found to be comparable to Butler (2003); although with a greater representation of oak (the relative availability of oak between the two studies is not known). Parakeet nests were recorded in trees of twelve genera, with oak accounting for 74%. The mean height of parakeet nest trees was 21.2 ± 5.97m with a DBH of 104 ± 40cm; the mean height of nests was 9.3 ± 3.3m. The majority of nests (77%) were in woodpecker cavities.

Some hole-nesting species exhibit a preference in the compass orientation of the nest cavity. In the present study, parakeets did not exhibit a preference for any orientation, with cavity entrances being uniformly distributed with compass bearing; also recorded by Butler (2003). Ring-necked parakeets, therefore, are not constrained in nest choice by the bearing of the cavity entrance. With respect to this parameter, parakeets nest site ‘preference’ overlapped with that of all native species.

Previously, Strubbe and Matthysen’s (2007) study showed that nuthatch and starling display the greatest overlap with parakeets in terms of typical nest cavity characteristics. The findings of the present study were consistent with this observation, with nuthatch and starling showing fewer significant differences in habitat features with parakeets in comparison to great-spotted woodpecker and jackdaw. In the case of jackdaw, the results were also consistent with Strubbe & Matthysen’s (2007) conclusion that potential competition for nest sites with parakeets was relatively low due to the jackdaw’s preference for nesting in wider, natural cavities, also preferred by stock doves. Ring-necked parakeets and nuthatches exhibited the greatest mean heights for both nest tree and nest cavity which is consistent with previous observations that ring-necked parakeets tend to use the highest cavities available for nesting (Pithon 1998, Bluekens 2002 cited in Strubbe & Matthysen 2007), that are also preferred by nuthatches (Strubbe & Matthysssen 2007).

For other native species, sample sizes were too small for analysis. For green woodpecker, however, previous studies have shown that although ring-necked parakeet nest sites included old green woodpecker cavities, the cavities occupied tended to be higher (mean 9.4m - Butler 2003) than the mean height of green woodpecker cavities (mean 4.6m - Glue and Boswell 1994). The present study showed qualitatively similar relative mean values for cavity height (parakeet = 9.3m, green woodpecker = 6.2m). Also, values for green woodpecker were almost identical to those for great-spotted woodpecker (Table 7.6), which were significantly different to parakeet.

Despite the overall frequent overlap in nest habitat characteristics between parakeets and native species and hence potential for competition for cavities, there was no apparent effect of parakeet abundance on nest site characteristics of native species. An effect of parakeets on native species would be most likely detected between sites exhibiting the greatest variation in parakeet abundance. However, when comparing the nest site habitat features of individual native species between sites with zero/low parakeet abundance and sites with high parakeet abundance there was an almost complete absence of any significant differences. Further, some of the few differences observed were in the opposite direction to that expected. Nuthatch and starling, for example, occupied taller trees in sites with high parakeet abundance, when any competitive exclusion would intuitively be expected to reduce nuthatches preferred use of the tallest trees (parakeets also prefer the tallest trees). One observation, however, that was consistent with potential competition for nest sites was that starlings and great-spotted woodpecker had a higher proportion of nests in branches (rather than trunks) in sites with high parakeet abundance. This is what would be predicted if parakeets were selecting cavities in the trunks and other species were therefore using a bigger fraction of cavities in branches when parakeet density was high. The observation of a greater proportion of parakeet nests in branches in high density parakeet sites is also consistent with this interpretation.

Of particular relevance to the findings was the markedly high abundance of tree cavities, a key habitat feature that would play a critical role in any competition for nest sites between parakeets and native species. The range of cavity densities of 22-94 cavities (>4<8cm entrance diameter) per ha was markedly greater than that recorded by Strubbe & Matthyssen (2007) in Belgium (8-18 cavities per ha). Indeed, 13 of the 15 sites in which habitat data was measured had a cavity density at least two-fold greater (2.1-5.2) than that of the site with the highest cavity density (18 cavities per ha) in Belgium. The method of counting cavities was similar in the two studies, with 0.05ha circular-sample plots searched from the ground for a standard time; this period was actually shorter in the present study (5 minutes per plot) compared to Strubbe & Matthyssen’s (2007) study (10 minutes). At these densities it is unlikely that cavities are a limiting factor in the Greater London area, at the present time, a conclusion previously made by Pithon & Dytham (1999). In light of the super-abundance of potential nest sites, the lack of evidence for an effect of parakeet abundance on nest site selection by native species, in the present study, is not surprising and is consistent with existing knowledge on relationships between nesting density and the availability of suitable cavities (Newton 1994).

SID 5 (Rev. 07/10) Page 19 of 37

8. Breeding parameters

8.1 Breeding parameters As stated in section 7.1, the potential consequences of competitive exclusion by parakeets of native species and displacement of the latter into less preferred nest cavities may be delayed or abandoned breeding. The potential consequences of the former may include reduced breeding success (fewer offspring fledged). The occurrence of an impact on the breeding success of native species was investigated by comparing the timing and productivity of breeding by native species between groups of sites with the lowest and highest abundance of parakeets respectively - Bedford to Wimbledon and Greenwich to Richmond, respectively (see Figure 6.1).

8.2 Methodology For all nest cavities that were accessible, multiple visits were used to determine information on breeding efforts. Monitoring of nest contents was achieved using a pole-mounted video camera.

The camera system comprised a miniature camera (2.3cm [l] x 1.5cm [b] x 1.7cm [h)]) fitted with LEDs attached to a 30cm flexible arm. The camera with AV transmitter was screw-mounted to the top of a carbon-fibre extendable pole (45ft full extension) such that the flexible arm extended perpendicular to the pole, which following extension of the pole to the appropriate height allowed insertion of the camera into a tree cavity. Images were transmitted to a hand-held viewer/recorder and stored on a memory card.

There were no consistent limitations in the ability of the equipment to access tree cavities and to capture images of the contents. On a small number of occasions the extensive nature of a cavity interior constrained a complete survey of its interior. Of the 235 nests discovered, only 12 of five different species were located higher than the camera pole’s reach.

Breeding parameters investigated were first-egg date (day on which the first egg of a clutch was laid) and fledgling success.

For each nest, the first-egg date was estimated by back-calculation from the observed status of the nest contents during nest monitoring visits, with reference to published data on clutch size, incubation period and nestling period for the target species (e.g. www.bto.org/birdfacts ). This process included, for example, back-calculating from the initial observation of a partial clutch, from the estimated age of nestlings, or from the assumed date of fledging (taken to be the date half way between the final record of the presence of mature nestlings and an empty nest cavity). The median first-egg date represented the date by which 50% of the target population had initiated egg-laying.

First-egg dates were allocated a numerical value corresponding to the day of the year (January 1 st = 1); e.g. a first-egg date of April 1 st equalled day 91. First egg dates for native species were compared between the two groups of sites with the most extreme difference in parakeet abundance (i.e. zero/low and high parakeet abundance) using Mann- Whitney U-tests.

The number of young successfully fledged from a nest was taken as the number observed to be present during a nest visit relatively late in the nestling period – for all nests this was >50% of the mean nestling period and for over 80% of nests this was >75% of the incubation period. The number of young fledged was compared between the two groups of sites with the most extreme difference in parakeet abundance (i.e. zero/low and high parakeet abundance) using Mann-Whitney U-tests.

8.3 Results

8.3.1 First-egg date

Ring-necked parakeet Of 62 ring-necked parakeet nests located, it was possible to estimate the first-egg date for 37 nests. The earliest that eggs were estimated to be laid was 8 th March. Median first-egg date was 24 th March. The last recorded first-egg date was 25 th April.

Native species Data on first-egg date was collected on a total of 119 nests of 8 native species (Table 8.1). Median first-egg date for ring-necked parakeet was a minimum of 18 days earlier compared to any native species (range 18 days [blue tit] to 45 days [green woodpecker]). For six native species sample sizes were large enough to test for a difference in first-egg date between: (i) native species and ring-necked parakeets, and (ii) sites with zero/low parakeet abundance and sites with high parakeet abundance.

First-egg date for ring-necked parakeets was significantly earlier than for all six native species (Mann-Whitney U-Tests P<0.001; sample sizes in Table 8.1).

SID 5 (Rev. 07/10) Page 20 of 37

Table 8.1. Median first-egg date for each species in sites grouped by relative parakeet abundance.

Parakeet abundance Species Zero/Low High All sites a Stock dove 27 April (3) 21 April (2) 25 th April (5) Ring-necked parakeet 24 th March ( 37 ) Green woodpecker 5 May (1) 8 May (5) 8 May (7) Great spotted woodpecker 22 April (20) 24 April (5) 22 April (27) Jackdaw 16 April (25) 15 April (8) 16 April (43) Blue tit 11 April (1) no data 11 th April (1) Great tit no data 24 April (3) 24 th April (3) Coal tit no data no data no data Nuthatch 16 April (9) 15 April (3) 16 April (13) Starling 16 April (11) 28 April (9) 17 April (20) Median first-egg date = date by which 50% of the nesting pairs have initiated egg-laying. aAll sites includes those with intermediate parakeet abundance. Values in parentheses represent the number of nesting pairs.

For four native species sample sizes were large enough to test for a difference in first-egg date between sites with zero/low parakeet abundance and sites with high parakeet abundance. For starlings, first-egg date was significantly later (12 days) in sites with high parakeet abundance (Mann-Whitney U-Test: U 9,11 =9.5, p = 0.02). There was no difference in first-egg date between sites with zero/low parakeet abundance and sites with high parakeet abundance for great spotted woodpecker, nuthatch or jackdaw (p>0.05).

8.3.2 Fledgling success

Ring-necked parakeets Data on fledgling productivity was collected on 37 ring-necked parakeet nests. The estimated mean number of offspring fledged was 2.3 ± 0.23 for all nests (n=37) and 2.7 ± 0.20 for successful nests only (n=31).

Native species Data on fledgling productivity was collected on 115 nests of 8 native species. The mean number of young fledged per nest is summarised in table 8.2. Fledgling success was compared between sites with zero/low parakeet abundance and high parakeet abundance (see section 6.2) - there was only sufficient data to compare fledgling success for four native species: great spotted woodpecker, jackdaw, nuthatch and starling. Comparisons were made for all nests and for successful nests only.

Table 8.2 Comparison of native species mean fledging rates (mean ± se; sample size in parentheses) between sites with zero/low and high parakeet abundance.

Parakeet abundance Zero/Low High All sites a Species All nests Successful All nests Successful All nests Successful nests nests nests Stock dove 0.5 ± 0.50 (2) 1.0 (1) 1.5 ± 0.50 (2) 1.5 ± 0.50 (2) 1.0 ± 0.41 (4) 1.3 ± 0.33 (3) Green woodpecker 5.0 (1) 5.0 (1) 4.0 ± 0.58 (3) 4.0 ± 0.58 (3) 4.0 ± 0.45 (5) 4.0 ± 0.45 (5) Great spotted woodpecker 4.6 ± 0.35 (18) 4.6 ± 0.35 (18) 3.6 ± 0.75 (5) 3.6 ± 0.75 (5) 4.3 ± 0.30 (25) 4.3 ± 0.30 (25) Jackdaw 1.9 ± 0.25 (31) 2.6 ± 0.18 (23) 2.4 ± 0.50 (9) 2.8 ± 0.45 (8) 2.0 ± 0.19 (50) 2.50 ± 0.15 (40) Blue tit 5.0 (1) 5.0 (1) no data no data 5 (1) 5 (1) Great tit no data no data 3.7 ± 1.45 (3) 3.7 ± 1.45 (3) 3.7 ± 1.45 (3) 3.7 ± 1.45 (3) Nuthatch 5.6 ± 0.84 (7) 5.6 ± 0.84 (7) 5.7 ± 1.45 (3) 5.7 ± 1.45 (3) 5.6 ± 0.62 (11) 5.6 ± 0.62 (11) Starling 3.2 ± 0.56 (8) 3.7 ± 0.36 (7) 2.9 ± 0.67 (7) 3.3 ± 0.56 (6) 3.1 ± 0.42 (15) 3.5 ± 0.31 (13) aAll sites includes those with intermediate parakeet abundance

There was no significant difference, for any native species, in the number of young fledged between sites with zero/low parakeet abundance and sites with high parakeet abundance, for both the number of young fledged per pair (all pairs) and for the number of young fledged per successful pair (Mann-Whitney U-Tests p>0.05).

There was no significant difference in nest failure rates between zero/low and high parakeet abundance sites: great spotted woodpecker 0% and 0%, jackdaw 25.8% and 11.1%, nuthatch 0% and 0% and starling 12.5% and 14.3% (zero/low and high parakeet abundance sites, respectively) (Fisher Exact, p>0.05).

8.4 Discussion Previous research showed that egg-laying dates of ring-necked parakeets in England are early compared to other native cavity-nesting species (Butler 2003). The first recorded date for a rose-ringed parakeet’s egg was between 27 th February and 5 th March, with a median first-egg date of 23 rd March (n = 108) and eggs being laid until mid May (Butler 2003). Ring-necked parakeets, on average, occupy nest cavities for a total of nine to ten weeks – three weeks incubation and six to seven weeks during the nestling stage (Forshaw 1989, Juniper & Parr 1998).

SID 5 (Rev. 07/10) Page 21 of 37 In the present study, the first parakeet egg was estimated to have been laid around the 8 th March (n=37) with the median first-egg date on 24 th March. The median fledgling date was estimated to be around 3 rd June. Parakeets, therefore, on average, occupied their nest cavities from around 24 th March to around 3 rd June. This period encompassed the egg-laying stage of native cavity nesters monitored (Cramp 1985, Cramp & Perrins 1993, 1994) and consequently could result in competition for nesting sites. Median egg-laying date in the ring-necked parakeet was significantly earlier than that of any of the native species (by 18-45 days). Median egg-laying dates for native species, however, fell within the published natural range for each species (Cramp 1985, Cramp & Perrins 1993, 1994).

The estimated mean number of parakeets fledged per nesting attempt was 2.2 ± 0.20 (n=37). This is higher than found in earlier studies in England: 0.8 chicks per nest (n=12) (Pithon & Dytham 1999) and 1.9 ± 0.1 chicks per nest (n=108) (Butler 2003). Nest failure by parakeets, due to nest abandonment, predation or chick death, was 11.3% (7/62 nests) across all sites.

There was no significant difference in the number of young fledged by native species (great-spotted woodpecker, jackdaw, nuthatch and starling) between sites with zero/low parakeet abundance and sites with high parakeet abundance. Similarly, the rate of nest failure of native species did not differ between sites with zero/low and high parakeet abundance. The lack of a difference in the number of young fledged or in the nest failure rate between zero/low and high parakeet abundance sites suggests that there are no differences in the quality of cavities used by native species in either category of site. Differences in nest characteristics that are consistent with variation in cavity quality have been shown to affect breeding success and/or productivity (Wesolowski 2003, Wesolowski & Rowinski 2004).

Starlings laid their clutches significantly later in sites with high parakeet abundance compared to sites with zero/low parakeet abundance - median first-egg date was 12 days later. This later clutch initiation, however, did not have an effect on breeding productivity (high parakeet sites = 2.9±0.67 fledglings per nest, zero/low parakeets sites = 3.2±0.56). However, it is known from previous avian studies that fledgling weight and subsequent survival is frequently lower in later-laid broods (e.g. Ingold 1996). In the present study, it cannot be known whether the later fledging dates in high parakeet abundance sites had any such detrimental effects post-fledging.

The results of the comparison of breeding parameters should be interpreted with caution. Sample sizes for most native species were either low, or very low which would limit the power to detect any significant differences.

9. Spatial analysis With the ring-necked parakeet population in England continuing to increase in numbers and spread geographically, it is of benefit to model the potential speed and pattern of this expansion. Such modelling will help identify variation in the spatial and temporal overlap between parakeets and populations of different native species. Irrespective of the level of any present competition between parakeets and native species, the magnitude of competition may increase as new areas are colonised by parakeets.

9.1 Methodology A preliminary investigation of the extent of potential future conflicts between an expanding ring-necked parakeet population and native cavity-nesting birds in England was undertaken using spatial analysis (ArcGIS software). The analysis involved four phases:

1) calculation of the edge of the current parakeet distribution, to act as the start point for scenarios of future spread, 2) analysis of historical data on the distribution of parakeets enabling the calculation of rates of population spread, 3) production of maps describing the density and distribution of potentially vulnerable native species across the study extent, and 4) intersection of these maps to permit the modelling of a series of simple scenarios investigating the spread of parakeets from their current range and the successive exposure of native species to potential competition over the next 30 years.

Data, scope and approach The rationale for choosing the data to model the past and current populations of ring-necked parakeets was fully detailed in a previous similar modelling exercise presented in Defra WM0104. This involved the use of BTO BirdTrack data of sightings of parakeets. Briefly, observations of birds submitted under the BTO BirdTrack online system were provided by the BTO/RSPB/BWI/SOC. This comprised 11,118 observations of ring-necked parakeets between 2003 and September 2010, each with a date, location and count of bird numbers. The observations are largely based on the BTO ’tetrad’ spatial reference, permitting the majority of points to be mapped to the nearest 2km. Some points required additional processing to improve the more limited spatial description supplied and to ensure it conformed to the tetrad approach. This usually involved enhancing the precision of the submitted co-ordinates (usually describing only the 10km grid square the sighting fell within) with whatever additional information was available in the ‘notes’ data field. Only five records had to be removed due to failure to derive or assume with confidence a suitably precise spatial description.

SID 5 (Rev. 07/10) Page 22 of 37

The data used to describe the density and distribution of other cavity nesting species were obtained from the BTO. For most species this represented the estimate of the absolute regional densities following the approach used in Newson et al. (2008). These densities were derived from the more standardised BTO/JNCC/RSPB Breeding Bird Survey (BBS). The regional densities of one species, the lesser spotted woodpecker, could not be estimated using this approach and as an alternative, an approximate description of the relative spatial distribution of this species using BirdTrack data was produced. Ideally, BBS data would be used for all species, but the random sampling design of the BBS means that it is less able to pick up scarcer species at current levels of coverage.

The geographic extent of the model study area was chosen to be within 135km of the 2009 core population of parakeets, excluding the BTO region “STFS” (Staffordshire south). This represents all of southeastern England, and stretches west of Bristol, north of central Birmingham and almost as far north as Nottingham (Figure 9.1).

Current distribution of rose-ringed parakeets The approach largely followed that which was developed for Defra WM0104, and is fully described and discussed in that project report. In brief, the observations for 2009 (the last full calendar year of data available) were first aggregated temporally to determine the monthly maximum count for every location, and then aggregated spatially into a regular hexagonal grid (of cell area 153.92 km 2) to represent the activity of a local community of parakeets. The spatial scale was chosen to ensure that observations in adjacent cells could be assumed to be independent. Twenty alternative representations of the data were generated using grids of differing origin and orientation, a map of a mean representation was produced and a threshold of 24 observations of parakeets was selected to describe the outer edge of the annual ‘core’ populations. An activity threshold of 24 was selected as the minimum representation of a significant density or population of parakeets in a grid cell; either the continual monthly report of a pair of parakeets or the sighting of a flock of 24 birds. The description of the core population produced was used to represent the current core distribution of ring-necked parakeet populations across England.

Modelling of parakeet annual core areas used parakeet data from full calendar years. Current evidence did not suggest that parakeets undertake any seasonal movements. The aggregation of observations across appropriately scaled cells permitted use of the more substantial and powerful data for a complete calendar year, rather than limiting the model to using sightings across any smaller period. In addition, unlike some species, the breeding season for parakeets is a considerable part of a full year, with the first birds beginning in January (when birds are prospecting for cavities) and ending in May/June (when the last young are fledged).

Rate of spread Six additional annual core areas were calculated for years 2003 to 2008. From these, the rates of spread were calculated for the three established populations of ring-necked parakeets in England; centred on London, Margate and Manchester. For London, fifteen vectors (straight lines perpendicular to the leading edge of the spread) were chosen and the rates of spread along these vectors between 2003 and 2009 were measured and calculated. The vectors were chosen to encompass the full range of modes of population spread (e.g. uniform and punctuated). The vectors were used to model three scenarios representing the full range of population expansion – ‘fast’, ‘slow’ and ‘average’. The vectors representing the five net fastest rates of spread were chosen to calculate a mean ‘fast’ rate of spread. The five vectors representing the net slowest rates of spread were chosen to calculate a mean ‘slow’ rate. All fifteen vectors were used to calculate an overall ‘average’ rate of spread. The slope of the linear regression of annual distance travelled against time was used to determine the mean rate of spread for each scenario. Seven appropriately placed vectors were used to measure the rate of spread for each of the Margate and Manchester populations.

Density of native cavity-nesting species The densities of the ten hole-nesting native species considered in this study were derived from two sources. An estimate of the absolute regional density of nine of the ten species was provided by the BTO (based on BBS data) including; blue tit, coal tit, great tit, greater spotted woodpecker, green woodpecker, nuthatch, starling, stock dove and jackdaw. These were described as individual birds km -2 across 39 previously described BTO regions, and should be considered as a gold standard estimate of the distribution of these common and widespread species. It was assumed that birds were distributed evenly across each BTO region. Regions within the study extent where no data was available were calculated to have the mean of the densities in their contiguous neighbours. The lesser-spotted woodpecker could not have its distribution described in the same way as it is relatively rare and data sufficiently scarce to make the complex numerical approach used by Newson et al . (2008) inappropriate. However, the very scarcity of this species makes the evaluation of its potential for competition with parakeets all the more important. As an alternative the following approach was used to derive a relative distribution of lesser-spotted woodpeckers across the study extent.

SID 5 (Rev. 07/10) Page 23 of 37 The distribution of lesser-spotted woodpeckers within BirdTrack was provided by the BTO for the years 2007-2009 inclusive. The data was initially prepared in a manner similar to that used for parakeets, in that locations were all presented to a similar level of spatial accuracy and annual maximum counts per location derived. It was determined that each of these point sightings could be considered independent, as the minimum distance between points (2km) was greater than the distance a typical woodpecker is likely to fly within its annual home range (Wiklander, 1998; Wiklander et al. , 2000; Wiklander et al. , 2001). The maximum numbers of sightings at each point (invariably 1 or 2) were then aggregated into 10km grid squares. However, the relative abundance of woodpeckers seen is also a function of the birdwatcher effort used to create the sightings. To produce an approximate index of birdwatcher effort within BirdTrack, estimates were obtained of sightings, distributed across a 10km grid, for four widespread and common bird species included in the scheme; blue tit, dunnock, starling and blackbird. These were combined to produce a crude quantitative map of birdwatcher effort. A map of the relative index of woodpecker abundance was calculated by simply dividing the count of birds seen by the relative bird watching effort for each 10km grid square.

Exposure of target species to competition with parakeets The future spread of parakeets from their current distribution was modelled using three simple scenarios. The scenarios differed only in the assumed rate of spread; and used the estimates of spread calculated for ‘slow’, ‘fast’ and ‘average’ for the spread of parakeets out of London. The chosen rates were assumed to be constant (e.g. throughout the scenario) and spatially uniform and represent the best available information to inform the predictions. Predictions of where the parakeets may have reached were made for two-year intervals for the first 10 years, and five-year intervals up to 30 years into the future. The area of England covered by the core population of parakeets in any year was intersected onto the descriptions of native bird density Figure 9. 1. Map showing the st udy extent and the description and the numbers of birds in potential competition with of the annual core populations of ring-necked parakeets. parakeets calculated. This was expressed as a proportion of the total national population given in Newson . (2008), Figure 1 (a) shows the descriptions of the population cores for 2009 (thick for each native cavity nesting species except the lesser black lines) along with the study extent (thin black perimeter line) and the spotted woodpecker, for which a total index of adjusted BTO regions. Figure 1(b) shows the population cores for London and Margate for 2004 to 2009 along with the vectors used to measure the rate sightings across the UK to represent the national of spread. population was produced. The modelling assumed that the distributions of native species remained constant over time. This may be valid for most of the common and widespread species, but not for all, e.g. green woodpecker and nuthatch are expanding northwards.

9.2 Results Current distribution of rose-ringed parakeets Ring-necked parakeet core areas for 2004-2008 were similar or identical to those calculated for Defra WM0104 (Figure 9.1). The core for 2009 suggests that as well as the continued expansion of the parakeet population from the cores observed in previous years (i.e. London, Margate and Manchester) that sufficient observations from six other locations have been made to suggest further populations are beginning to establish (Figure 9.1a). Two of these may be locations associated with very frequent bird-watching activity (e.g. Dungeness and the Ribble estuary) and may be artefacts of data uncontrolled for observational effort. However, one location is centred on central Birmingham and follows the pattern of previous establishments of ring-necked parakeets in being located in a major urban centre. Two of the locations are along the southern coast of England and may be the results of spread from the Margate population. The veracity of these potential new establishing populations requires ground-truthing.

Rate of spread The rates of spread of parakeets were calculated between the year 2004 and 2009. The mean rate of spread of parakeets out of London was 2.89km year -1. There is considerable variation in both annual spread and the overall distance moved when considering each of 15 individual vectors (Figure 9.2) and the temporal pattern of spread is erratic. For example, examining vector 4 demonstrates a very slow net rate of spread characterised by an initial retreat SID 5 (Rev. 07/10) Page 24 of 37 of approximately 3km from its position in 2004, very little movement for four years and a substantial 6km jump in the final year of study. In contrast the rate of spread along vector 1 is usually positive, with annual movements ranging between 0km (in 2007) up to 11.8km (in 2005), and representing a net advance of the population core of 27.2km over the five years in this study. The ‘slow’ and ‘fast’ rates are 1.3km year -1 and 4.5km year -1 respectively (Figure 9.3). All three rates are in a similar range to those found in WM0104.

It is notable that even vectors with slow net rates of progress can show significant annual advances. Examining all 15 vectors and considering only the largest single annual advance in each, we note that maximum annual advances range between 4km and 13km with a mean of 8.13km and a standard deviation of 2.8. The significance of this is it suggests that parakeets have the potential to make considerable annual advances across apparently inhospitable landscapes where mean rates are low.

The net rates for Margate and Manchester are also positive though are determined from much smaller populations (Figure 9.3). The Margate population shows a mean rate of spread of 1.2km year -1, though again examining the maximum annual rate yields a range of 4.9km to 14.7km with a mean maximum potential spread of 10.1km year -1. The existence of a potential new population in Manchester was first detected for the last year of study (2008) in project WM0104. In the present study this is re-confirmed as an established population in 2008 and is present in the model for 2009, giving only one year across which to measure an advance of the population front, though this is a substantial mean advance of 6.88km.

Figure 9.2. The mean rates of spread of ring-necked Figure 9.3. The annual spread of ring-necked parakeets parakeets for a range of contexts. along 15 vectors in London.

Density of native cavity-nesting species The densities of nine of the 10 native species were mapped directly from data provided by the BTO (based on BBS data). Many species show distinct patterns in their regional densities; for example nuthatch (Figure 9.4a) and green woodpecker (Figure 9.4b) show western and south-eastern biases in density respectively. Starling (Figure 9.4c) is also presented as this is a widespread species of interest, broadly associated with man, and was the only species to show variation in a breeding parameter associated with a difference in the abundance of ring-necked parakeets (see section 8).

Despite the availability of sighting data for three full years, the distribution of the lesser-spotted woodpecker that it describes is inevitably discontinuous as this species is small, scarce and easily over-looked.

The assumption was made that BirdTrack sightings were independent and the maximum number of birds seen at any location in any year were directly summed into a suitable geographical format, in this case the same 10km grid used by the BTO for their other data. The map of bird watcher effort, whilst very approximate does confirm that across the study extent, bird watching effort and subsequent submission of records to BirdTrack was continuous and substantial compared to other regions of England.

A single map of the number of sightings per 10km square, adjusted for the bird-watching effort, was produced to represent the relative density of lesser-spotted woodpeckers across the study extent (Figure 9.4d).

SID 5 (Rev. 07/10) Page 25 of 37

(a) (b)

(c) (d)

Figure 9.4 Maps of the densities of native cavity nesting birds across the study extent: (a) absolute density of nuthatch (b) absolute density of green woodpecker (c) absolute density of starling (d) relative density of lesser spotter woodpecker

SID 5 (Rev. 07/10) Page 26 of 37 WM0104 Annex to Sid5

Exposure of target species to competition with parakeets

Individual species of cavity-nesting birds differed in their exposure to potential competition with ring-necked parakeets. The estimated rates of spread used to model the exposure were 1.3 km year -1, 2.9 km year -1 and 4.5 km year -1. The differences in species’ exposure are due to the different distributions of the cavity-nesters and how parakeets invade these. For example, the species predicted to be potentially most exposed to the continued spread of parakeets is the green woodpecker, with up to 79% of the national population exposed to potential competition with parakeets after 30 years under the scenario of most rapid population spread. Even considering the mean scenario of spread at 2.9 km year -1, 31% of green woodpeckers will share landscapes with parakeets after only 10 years.

Other species are less affected and their presentation in figure 9.5 shows them in decreasing rank order based on the proportion of their national population exposed to parakeets over a 30-year period (under the mean scenario). The comparative ranks of proportional exposure after 0, 10 and 30 years under the mean scenario can be seen to vary (Table 9.1), due to the differential regional distributions of the native species. Some species, such as the lesser-spotted woodpecker may already be considered to be considerably exposed to potential competition with parakeets, with almost 20% of the national population presently sharing breeding sites.

The rank order of species shows some variation when ‘slow’ and ‘fast’ rates of parakeet spread are used to model the exposure of national populations.

Table 9.1 Ranked modelled exposure of cavity-nesting birds to parakeets for 2009, and in 10 years and 30 years, under the mean rate of spread.

Percentage exposure to parakeets Species Year 2009 2019 2039 Figure 9.5 Modelled exposure of native cavity-nesting Green woodpecker 10.0 31.0 57.0 birds to parakeets under a range of scenarios of spread Lesser spotted woodpecker 18.6 33.2 48.8 Greater spotted woodpecker 7.7 21.1 37.4 - Fast rate , mean rate and slow rate . The y axis shows Stock dove 5.3 17.1 36.2 the proportion of the national population exposed - note Nuthatch 5.6 17.2 35.2 variation in axis scale. Great tit 4.4 13.4 27.9 Blue tit 5.1 14.7 27.8 Starling 6.6 15.7 25.9 Jackdaw 2.1 7.4 18.8 Coal tit 1.6 5.0 12.2

27

WM0104 Annex to Sid5

10. Overview

Potential environmental impact of parakeets Introduced non-native birds can compete with native species for various resources, including nest sites. One example is the European starling in the USA. A number of studies have documented competition between starlings and native cavity-nesters. In Arizona, Kerpez & Smith (1990) found that starlings competed with native gila woodpeckers Melanerpes uropygialis for nest cavities and resulted in a negative relationship between the numbers of starling and gila woodpecker nests. In Ohio, northern flickers were found to frequently lose nest cavities to European starlings, and although flickers routinely renested later in the season, late nesting pairs fledged fewer offspring than early nesting pairs (Ingold 1966, 1998). Conversely, other studies have failed to demonstrate a detrimental effect of starlings on native species, despite clear examples of competition for cavities (Troetschler 1976, Koenig 2003).

Kimschick and Nentwig (2010) produced a scoring system for introduced non-native birds that compared their impacts (potential and actual) in Europe. The ring-necked parakeet was ranked third (after Canada goose Branta canadenis and ruddy duck Oxyura jamaicensis ) in terms of actual environmental impact and second (after Canada goose) in terms of the highest actual economic impact score.

There is also some evidence for an ecological impact of non-native ring-necked parakeets on native cavity- nesters. Competitive exclusion of native species from preferred cavities by parakeets could occur through through aggressive displacement or usurpation...In addition, because they initiate breeding significantly earlier than native species (in the present study, 18-45 days dependent on species), parakeets are able to colonise the most preferred and high quality cavities. In Belgium, Strubbe & Matthysen (2007) explored the relationship between the numbers of ring-necked parakeets and native species and habitat and landscape variables. A negative association between parakeet and nuthatch numbers was concluded to potentially indicate competition for nesting cavities. More recently, Strubbe and Matthysen (2009) conducted a replicated before-and-after field experiment in which parakeet nesting cavities were blocked. Following blocking, numbers of nuthatches declined, largely due to nest take-overs (through occupancy earlier in the season) by parakeets. Mean quality (as measured by previous usage) of parakeet nests decreased indicating that parakeets will utilise alternative cavities when preferred sites are not available, increasing the potential for conflict with native species.

In Britain, the continued population expansion of ring-necked parakeets has raised concern over potential detrimental impacts on native cavity-nesters here. The present study addressed this concern by investigating three issues: (i) identify the impact of ring-necked parakeets on the distribution and abundance of native species, (ii) quantify the likely significance of these impacts on local populations of native species, and (iii) assess the likely risk to the population status of native species under different scenarios for spread of the ring-necked parakeet.

Evidence of impacts in the UK This study provided no current evidence for a detrimental impact on the distribution and abundance of native species at the national scale. Analysis of the BBS dataset indicated that neither the numbers of native cavity nesters nor their population growth rates demonstrated any significant negative correlation with the abundance of parakeets. While it is possible that parakeets may still have a negative impact at a subset of sites where the competition for cavities is high, a further analysis of intensive monitoring data at 20 sites failed to find any overall relationship between the abundance of parakeets and that of most native cavity nesters, even when those habitat characteristics likely to correlate with cavity availability were taken into account. There was a suggestion, however, of some degree of local-scale competition between parakeets and both green woodpecker and stock dove in sites with relatively low proportions of continuous closed canopy woodland.

While the characteristics of nest sites chosen by parakeets and those selected by some native cavity nesting species, did show overlap, there was little evidence that the nest sites of native species differed between sites with low parakeet abundance and sites with high parakeet abundance, as would be consistent with a general absence of displacement of native species into unsuitable nest cavities.

In addition, no evidence was provided for a detrimental effect on the breeding productivity of native species. The timing of breeding of native species was also unaffected, except for starling which exhibited a median first-egg date that was two weeks later in sites with high parakeet abundance, compared to sites with zero/low parakeet abundance. It is not known whether this delayed breeding had any subsequent consequences for the survival of young after fledging, recruitment to the population or population size and we cannot exclude the possibility that laying dates were influenced by habitat characteristics that were also correlated with parakeet abundance.

Factors mitigating strength of competition for nest sites The evidence for the negative influence of parakeet numbers in the Belgian population provided by Strubbe and Matthysen (2007, 2009) and the contrasting absence of a significant effect in the UK was initially surprising given the overlap in some nest site preferences demonstrated in the current study. The power of the statistical tests 28

WM0104 Annex to Sid5 relating numbers of native species to the abundance of parakeets is high, as indicated by the relatively small changes in blue tit population growth rates that were identified by analysis of the BBS dataset. It is therefore unlikely that significant relationships are present but were not detected. However, the density of tree cavities may be a key factor in mediating the impact of parakeets. Any competition for cavities with preferred characteristics is likely to be mitigated if alternative sites are readily available. Thus, although parakeets have the first opportunity to acquire high quality cavities through early breeding, opportunities for native species to do likewise, later in the season, may be relatively little reduced.

The median density of 55 cavities (>4<8cm entrance diameter) per ha was much greater than that recorded by Strubbe & Matthyssen (2007) in Belgium (8-18 cavities per ha) and by Wesolowski (1989) in European deciduous woods (5-17 cavities per ha). Indeed, 13 of the 15 sites in which habitat data were collected had a cavity density at least two-fold greater (2.1-5.2) than that of the site with the highest cavity density (18 cavities per ha) in Belgium. At these densities it is unlikely that cavities are a limiting factor in the Greater London area, at the present time, a conclusion previously made by Pithon & Dytham (1999). In light of the super-abundance of potential nest sites, therefore, the lack of evidence for an effect of parakeet abundance on nest site selection by native species, in the present study sites, is not surprising and is consistent with existing knowledge on relationships between nesting density and the availability of suitable nesting cavities (Newton 1994, Lohmus & Remm 2005). In both the present study and Butler (2003) the majority of ring-necked parakeet nests were located in cavities originally excavated by woodpeckers. Possible differences in the densities of woodpeckers (and hence cavity production) between England and Belgium is one potential explanation for the difference in cavity abundance.

Alternatively, the strength of competition for nest sites in the UK may be reduced because absolute densities of one or both of the competing species are lower. The different methodology used to assess the numbers of bird species in the Belgian study (point counts) and the current study (line transects) make it impossible to compare densities directly, but there is scope for future collaborative work which would enable this comparison.

In relation to the issue of competition with cavity-nesting species in general, the dynamics of cavity-nesting communities, have been shown to involve far more complex interactions than the availability of cavities per se. Some secondary-cavity nesting species may be limited by the availability of suitable cavities, whereas as other species may be limited by more dominant cavity-nesters (Aitken & Martin 2008). It was suggested that plasticity in nest site selection may enable some species to resist or respond positively to changes in cavity availability, whereas other species were suggested to be limited by more specialised nest site preferences.

Potential for future impacts of parakeet populations It should be stressed, however, that the absence of a detectable impact on the abundance and distribution of native cavity nesters does not imply a continued absence of impact in the future. At the local scale, the present study represents a snapshot of the relative abundances, nest site preferences and breeding performance of parakeets and native species in the Greater London parklands.

It has been argued that the total impact of an invasive species includes three fundamental dimensions: range, abundance, and the per-capita or per-biomass effect of the invader (Parker et al . 1999). The continued growth and geographical expansion of ring-necked parakeet populations in England satisfies increases in two of these three key dimensions. Kimschick and Nentwig (2010) also stated that the full extent of a species impact should be seen as an increasing component of its invasion history, as it enlarges its European range. The same concept can be applied when considering a species’ range expansion within an individual country.

The estimated rates of spread of parakeets out of London calculated in the current study were 1.3km year -1, 2.9km year -1 and 4.5km year -1, representing a ‘slow’, ‘average’ and ‘fast’ rate of spread respectively. The modelling showed that native species differed in the rate of exposure of their national populations to parakeet population expansion over the next 30 years. Under the mean estimated rate of spread of parakeets, the species that will become most exposed are green woodpecker (57% of the national population) and lesser-spotted woodpecker (49%). After only ten years, modelling indicates that over 30% of each of these species’ populations will occupy landscapes alongside parakeets. The least exposed are coal tit (12%) and jackdaw (19%). Of note, is that the species predicted to become the most exposed, the green woodpecker, is also the species for which there was some evidence for an impact of parakeets on their abundance at some local-scale sites. Also of note, is that the species predicted to become the second most exposed is the lesser spotted woodpecker, a species that has experienced significant declines in its abundance and distribution over the past 30 years. Consequently, any study of the decline in lesser spotted woodpecker should consider the potential role of any impacts due to ring- necked parakeets.

With further increases in the breeding density of parakeets in their current range and expansion into new areas the potential for competition for nest cavities is likely to increase. The availability of suitable nest cavities is, on average, likely to be lower outside of the London parks, used as study sites in the present study. The current hotspots for parakeets are in areas with large numbers of mature trees that are likely to be more common around 29

WM0104 Annex to Sid5 human habitation, such as the London Parks. Mature trees are less likely in agricultural land or more managed woodland (van Balen et al . 1982, Newton 1994). The lower availability of tree cavities in other areas in England is consistent with Pithon and Dytham’s (1999) study that showed parakeets utilising only 0.6% of nestboxes (n=175) in the Greater London area but 26% (n=23) on the Isle of Thanet, Kent.

Strubbe & Matthysen (2010) modelled the potential impact of ring-necked parakeets on nuthatches across Flanders. Their approach involved modelling the expected abundance of ring-necked parakeets and combined it with an empirically derived measure of competition strength to assess the magnitude of expected impact on nuthatches. It was considered that the total impact of parakeets would be limited with, in the most extreme scenario, about one third of the nuthatch population being lost. It was recognised that more severe impacts on some rare and threatened species (e.g. wryneck Jynx torquilla ) could not be excluded, in the event of further expansion of the parakeet population.

Strubbe & Matthysen’s (2010) approach assumed some relatively distinct differences in preference for nesting habitat between parakeets and nuthatches, with fine-scale separation in the landscape retained following the modelled spread of parakeets into all suitable sites. Such an assumption is not necessarily valid for England. As discussed above, the extent and magnitude of competition is likely to vary with numerous factors, such as habitat, composition of the cavity-nesting bird community and species’ behavioural flexibility.

In England there remains the potential for competition and detrimental impacts of parakeets on native species outside the present study area. Potentially, significant proportions of the national populations of a number of cavity-nesting species may become exposed to potential risk of competition for nest sites. Although modelling shows that some species are relatively more vulnerable than others due to a greater rate and magnitude of exposure of the national population, any impacts that may be manifest are likely to vary with habitat characteristics. For example, a native species with a greater relative proportion of its national population that is associated with more resource-limited habitat (e.g. lower density of suitable cavities) is more likely to be impacted upon than a species similarly occupying more resource-rich habitat.

Further data on factors, including potential impacts in existing cavity-limited sites, relative distributions of native species relative to cavity-availability and likely future parakeet nesting densities in these habitats, are required to enable a detailed evaluation of the risk to native species’ populations.

Scope of current study The findings of the study should be interpreted with consideration of the limitations in the scope of the investigation, imposed by the nature of the datasets available and the duration of the fieldwork being constrained to a single breeding season. The consequence of these constraints was that the investigation of the interactions between parakeets and native species were limited spatially and temporally. The limitation of study sites to parkland habitat was driven primarily by the distribution of nuthatch (the species avoids urban/suburban habitat), the only species for which a negative effect of parakeet abundance was identified in Belgium. However, although the study focussed on one habitat type this encompassed a range of parakeet densities including high density parakeet areas in which interactions with native species would be expected to first occur. Sample sizes were small in a number of parameters under investigation, particularly breeding parameters and, therefore, the power to detect any significant difference in these instances was low. A marked feature of the parkland sites was that cavity abundance was very high and possibly mitigated any negative effects of interactions. The study was not able to address the potential existence of current detrimental impacts in sites with a lower abundance of tree cavities. The present study devised some simple scenarios as to the potential nature of the future expansion of the ring-necked parakeet population and potential conflict with native cavity-nesting species. This spatial modelling was based on the best available data, but interpretation should be done with care as there is limited information on parakeet habitat use and spatial dynamics. One specific gap in knowledge is what factors influence variation in the rate of population expansion across its range and between years. An understanding of this, and development of more detailed spatial models, requires data on population dynamics, and on parakeet resource requirements and its availability and quality.

11. Conclusions Overall, the present study provides no consistent evidence that, within their current geographical range and at current densities in England, ring-necked parakeets are negatively influencing the abundance of any common native cavity nesting species at the national scale. There was a suggestion, however, of some degree of local- scale competition between parakeets and both green woodpecker and stock dove in sites with relatively low proportions of continuous closed canopy woodlands.

Despite identification of overlap in cavity preferences between parakeets and native cavity nesters, there was no evidence for any difference in the productivity of breeding attempts by native species (great-spotted woodpecker, jackdaw, nuthatch and starling) between sites with zero/low parakeet abundance and sites with high parakeet abundance. Starlings did commence egg-laying later in sites with high parakeet abundance but there was no 30

WM0104 Annex to Sid5 difference in the number of young fledged by starlings between the two categories of sites.

The abundance of tree cavities in the London study sites was very high; much higher than in the Belgium study sites that had previously recorded a detrimental effect of parakeet abundance on that of nuthatches. The absence of any detrimental effects in the present study is consistent with the higher abundance of tree cavities, but differences in the population densities of parakeets and/or native species may also influence the strength of nest site competition.

Although there was no firm evidence for any present detrimental effect of parakeets on native cavity nesters in their principal range, it is stressed that these findings do not exclude the possibility of conflict in the future; particularly as ring-necked parakeet populations are continuing to increase in numbers and expand geographically. The present overlap between the core areas of parakeets and native species’ populations at a national scale is, at present, low for most species. Spatial modelling, however, indicated that at the present rate of population spread, in thirty years parakeets are likely to occupy significant percentages (25-57%) of the national range of eight cavity-nesting species.

In situations where nest cavities are a more limited resource, ring-necked parakeets retain the potential to out- compete native species, potentially resulting in population declines of the latter. This also remains a possibility within their existing range as parakeet densities increase.

The scope of the study was limited by the nature of the datasets available, the duration of the fieldwork being constrained to a single breeding season and restricted to a single habitat type (parkland). Sample sizes were small in a number of parameters under investigation, particularly breeding parameters. Cavity abundance was very high and possibly mitigated any negative effects of interactions. The study was not able to address the potential existence of current detrimental impacts in sites with a lower abundance of cavities. Spatial modelling was based on the best available data, but interpretation should be done with care as there is limited information on parakeet habitat use and spatial dynamics.

12. Recommendations The present study has highlighted a number of gaps in knowledge for which further investigations are recommended:

The present study was limited to London parklands, for which a significant common feature is the abundance of large mature trees, and associated high density of cavities. Similar investigation of parakeet and native species abundance and breeding parameters would be beneficial in the other major region occupied by parakeets, i.e. Kent. Expansion of the Kent population has been slower than that of the Central London population. One potential reason for this may be a lower availability of suitable nest sites. Hence, the dynamics and consequences of parakeet-native species interactions may differ from that found in the present study. This is also the case for any newly establishing populations in other areas in England.

In light of the above, all locations (especially those in Manchester and Birmingham) identified as likely sites for newly establishing populations should be systematically surveyed to confirm the presence of ring-necked parakeets.

In the present study, as in all correlative studies, species and site effects are confounded and thus ascribing cause and effect is not possible. The strongest evidence for a detrimental effect of parakeets on native species would be provided through a field experiment that manipulated the abundance of parakeets and/or cavity availability. The manipulation of parakeet abundance would be very difficult to achieve and maintain. Cavity availability, however, could be manipulated in a cavity blocking experiment. An alternative approach to artificially manipulating parakeet abundance would be to undertake a natural experiment. This would involve monitoring the abundance and breeding efforts of native species in study sites at the edge of the parakeets expanding range, and also any changes in the species-occupancy of individual nest cavities as parakeets become progressively established in these areas.

Collaborative studies with the research group in Belgium would be very beneficial allowing, for example, complementary field experiments, or modelling to be undertaken in different parts of the ring-necked parakeet’s European range.

Refinement of the spatial modelling methodology would improve the authority and quality of the description of parakeet core areas. One approach would be to validate parakeet activity thresholds derived as part of the modelling process with a series of stratified random surveys (transects or point counts) in core parakeet areas and other areas with a range of parakeet colonisation history.

Further refinement of the spatial modelling would involve investigation of habitat use and potential seasonal 31

WM0104 Annex to Sid5 changes in the spatial dynamics of parakeets, as it may help determine one of the drivers of spread, and help describe one or more of the fundamental ecological requirements for the species (e.g. winter forage, nest holes, mates). Such investigation could be through the appropriate analysis of existing historical datasets and/or through targeted fieldwork (e.g. satellite tracking of parakeets).

32

WM0104 Annex to Sid5

Appendix I: Location of study sites.

33

WM0104 Annex to Sid5

References to published material 9. This section should be used to record links (hypertext links where possible) or references to other published material generated by, or relating to this project.

Newson, SE, Parrott, D., Leech, D. 2010. Evaluating the population level impact of an invasive species, Ring- necked Parakeet Psittacula krameri on native avifauna. Ibis (submitted manuscript). Annex 1.

34

WM0104 Annex to Sid5

Aitken, K.E.H. & Martin, K. 2008. Resource selection plasticity and community responses to experimental reduction of a critical resource. Ecology 89(4): 971-980. Brouwer, K., Jones ML., King CE. & Schifter H. 2000. Longevity records for Psittaciformes in captivity. International Zoo Yearbook 37: 299-316. Buckland, S.T. 2006 Point transect surveys for songbirds: robust methodologies. The Auk 123: 345-357. Butler, C. J. 2003. Population biology of the introduced roseringed parakeet Psittacula krameri in the UK. Unpublished Ph.D. thesis, Deptartment of Zoology, Edward Grey Institute of Field Ornithology, University of Oxford. Cramp, S. 1985. Handbook of the Birds of Europe, the Middle East and North Africa. The Birds of the Western Palearctic , Vol. IV. Terns to Woodpeckers. Oxford University Press, Oxford. Cramp, S. & Perrins, C.M. 1993. Handbook of the Birds of Europe, the Middle East and North Africa. The Birds of the Western Palearctic , Vol. VII. Old World Flycatchers to Shrikes. Oxford University Press, Oxford. Cramp, S. & Perrins, C.M. 1994. Handbook of the Birds of Europe, the Middle East and North Africa. The Birds of the Western Palearctic , Vol. VIII. Crows to Finches. Oxford University Press, Oxford. Defra WM0104. Rose-ringed parakeets in England: a scoping study of potential damage to agricultural interests and management measures. Fera report to Defra. Forshaw, JM. 1989. Parrots of the world . Blandford, London. Freeman, S.N. & Newson, S.E. 2008 On a log-linear approach to detecting ecological interactions in monitored populations. Ibis 150: 250-258. Glue, D.E. & Boswell, T. 1994. Comparative nesting ecology of the three British woodpeckers . British Birds 87: 253-269. Haines-Young R.H. et al. (2000) Accounting for nature: assessing habitats in the UK countryside. DETR, London. Ingold, D.J. 1996. Delayed nesting decreases reproductive success in Northern Flickers: implications for competition with European starlings. Journal of Field Ornithology 67: 321-326. Ingold, D.J. 1998. The influence of starlings on flicker reproduction when both naturally excavated cavities and artificial nest boxes are available. Wilson Bulletin 110(2): 218-225. James, F.C. & Shugart. H.H. 1970. A quantitative method of habitat description. Audubon Field Notes 24:727–736. Juniper, T. & Parr, M. 1998. Parrots: a guide to parrots of the world. Pica, Robertsbridge. Kerpez, T.A. & Smith, N.S. 1990. Competition between European starlings and native woodpeckers for nest cavities in saguaros. Auk 107: 367-375. Kumschick, S. & Nentwig, W. 2010. Some alien birds have as severe an impact as the most effectual mammals in Europe. Biological Conservation (in press). Koenig, W.D. 2003. European starlings and their effect on native cavity-nesters. Conservation Biology 17(4): 1134-1140. Lehmkuhl, J.F., Everett, R.L., Schellhaas, R., Ohlson, P., Keenum, D., Riesterer, H. & Spurbeck, D. 2003. Cavities in snags along a wildfire chronosequence in eastern Washington. Journal of Wildlife Management 67:219-228. Lohmus, A. & Remm, J. 2004. Nest quality limits the number of hole-nesting passerines in their natural cavity-rich habitat. Acta Oceologica 27(2): 125-128. Mazgajski, T.D. 2007. Nest hole age decreases nest site attractiveness for the European starling Sturnus vulgaris . Ornis Fennica 84: 32-38. Newson, S.E., Evans, K.L., Noble, D.G., Greenwood, J.J.D. & Gaston, K.J. 2008. Use of distance sampling to improve estimates of national population sizes for common and widespread breeding birds in the UK. Journal of Applied Ecology 45: 1330-1338. Newton, I. 1994. The role of nest sites in limiting the numbers of hole-nesting birds: a review. Biological Conservation 70: 265-276. Parker, I.M., Simberloff, D., Lonsdale, W.M., Goodell, K., Wonham, M., Kareiva, P.M., Williamson, M.H., Von Holle, B., Moyle, P.B., Byers, J.E. & Goldwasser, L. 1999. Impact: toward a framework for understanding the ecological effects of invaders. Biological Invasions 1: 3-19. Pell, A.S. & Tidemann, R . 1997 . The impact of two exotic hollow-nesting birds on two native parrots in savannah and woodland in eastern Australia . Biological Conservation 79 : 145 –153 . Phillips, S.J., Anderson, R.P. & Schapire, R.E. 2006 Maximum entropy modeling of species geographic distributions. Ecological Modelling 190: 231-259. Pithon J.A. 1998. Status and ecology of the Ring-necked Parakeet Psittacula krameri in Great Britain. PhD thesis, University of York. Pithon, J.A, & Dytham C. 1999. Breeding performance35 of Ring-necked Parakeets Psittacula krameri in small introduced populations in southeast England. Bird Study 46: 342-347.

WM0104 Annex to Sid5

Pithon, J.A, & Dytham C. 1999. Breeding performance of Ring-necked Parakeets Psittacula krameri in small introduced populations in southeast England. Bird Study 46: 342-347. SAS Institute. 2001. SAS/STAT User’s Guide, version 8.02. North Carolina: Cary. Strubbe, D. & Matthysen, E. 2007. Invasive ring-necked parakeets Psittacula krameri in Belgium: habitat selection and impact on native birds. Ecography 30: 578-588. Strubbe, D. & Matthysen, E. 2009. Experimental evidence for nest-site competition between invasive ring- necked parakeets ( Psittacula krameri ) and native nuthatches ( Sitta europaea ). Biological Conservation, 142: 1588-1594. Strubbe, D., Matthysen, E. & Graham, C.H. 2010. Assessing the potential impact of invasive ring-necked parakeets Psittacula krameri on native nuthatches Sitta europeae in Belgium. Journal of Applied Ecology 47(3): 549-557. Troetschler, R.G. 1976. Acorn woodpecker breeding strategy as affected by starling nest-hole competition. The Condor 78:151-165. van Balen, J.H., Booy, C.J.H., van Franeker, J.A. & Osieck, E.R. 1982. Studies on hole-nesting birds in natural nest sites 1. Availability and occupation of natural nest sites. Ardea 70:1-24. Wesolowski, T. 1989. Nest-sites of hole-nesters in a temperate forest (Bialowieza National Park, Poland). Acto Ornithologica 25(3): 321-349. Wesolowski, T. 2003. Clutch size and breeding performance of Marsh Tits Parus palustris in relation to hole size in a primeval forest. Acta Ornithologica 38(1): 67-74. Wesolowski, T. & Rowinski, P. 2004. Breeding behaviour of Nuthatch Sitta europaea in relation to natural hole attributes in a primeval forest. Bird Study 51: 143-255. Wiklander, U., Nilsson, S.G. & Olsson, O. 2000. Parental care and social mating system in the Lesser Spotted Woodpecker ( Dendrocopos minor ). Journal of Avian Biology 31:447-456. http://onlinelibrary.wiley.com/doi/10.1034/j.1600-048X.2000.310003.x/pdf Wiklander, U., Olsson, O. & Nilsson, S.G. 2001. Seasonal Variation in home-range size, and habitat area of the lesser spotted woodpecker ( Dendrocopos minor ) in southern Sweden. Biological Conservation 100(3):387-396. Wiklander, U. 1998. Reproduction and survival in the lesser spotted woodpecker. Effect of life history, mating system and age. ISBN: 91-7105-096-5. http://swepub.kb.se/bib/swepub:oai:lup.lub.lu.se:38487?tab2=abs&language=en

36

WM0104 Annex to Sid5

Acknowledgements We are very grateful to the following members of Fera staff: Christopher Rhodes, Sara Bone and Paul Cropper for fieldwork, Stephan Pietravalle for statistical services and Nathalie Key for assistance with GIS analysis, and BTO staff: Bill Haines for undertaking the bird and habitat survey fieldwork in the London parks; Greg Conway, Graham Austin and Maria Knight for their GIS work to determine habitat coverage of the London parks; Alison Johnston for statistical advice; Mark Grantham, Lee Barber and Liz Coiffait for undertaking pilot fieldwork; John Tayleur for providing background information.

We would like to thank all the landowners and individuals who enabled access to the fieldwork sites: Richmond Park, Bushy Park, Hyde Park, Kensington Gardens, Green Park, St James' Park, Greenwich Park and Regent’s Park - Nigel Reeve, Simon Richards, Ray Brodie, Steve Edwards, Mark Wasilewski, Tom Jarvis, Derrick Spurr and Nick Biddle at The Royal Parks; Windsor Great Park - Bill Cathcart at The Crown Estate; Kew Gardens - Steve Ruddy, Sandra Bell and Gavin Kinsella, at which access was granted by permission of the Trustees of the Royal Botanic Gardens, Kew; Knole Park – Crispin Scott, Steve Dedman, Valerie Porter and Robert Broughton at National Trust, Phil Williams at Natural England and Isabell Swift at Strutt & Parker; North Mymms Park - Johnny Carew-Pole at North Mymms Park, Debbie Youngs and Steph Cato at GlaxoSmithKline, Peter Jones at Colney Bridge Farm; Langley Park - Tim Williams and Matt May at Buckinghamshire County Council; Peter Haldane and Liz Duvar at Wimbledon & Putney Commons; Hampstead Heath – Adrian Brooker at City of London Corporation and Mike Taylor at English Heritage; Cassiobury Park – Ian Mather at Cassiobury Park, Ross McCue at West Herts Golf Club; Osterley Park – Jeremy Dalton and Alice Springate at National Trust; Silwood Park – Mick Crawley and Alex Lord at Imperial College, London; Lamberhurst Park – Ross Wingfield at National Trust; Bedford’s Park – Geoff Pepper at Havering Council; Weald Country Park – Claire Menim and Nicole Khan at Essex County Council; Grovelands Park – Trevor Richards and Leila Briscow at Enfield Council, and Chris Morris; Cliveden Gardens – Alex Livingstone at National Trust.

We also thank the thousands of volunteers involved in collecting information for the suite of BTO surveys that provided data for the analyses presented in this report, and the organisations involved in the development and funding of this work. The Breeding Bird Survey is run by the BTO on behalf of the BTO, the Joint Nature Conservation Committee (JNCC) (on behalf of the statutory nature conservation agencies: Council for Nature Conservation and the Countryside, the Countryside Council for Wales, Natural England and Scottish Natural Heritage), and the Royal Society for the Protection of Birds (RSPB). BirdTrack is organised by the BTO on behalf of the BirdTrack partners (BTO, RSPB, BirdWatch Ireland, Scottish Ornithologists’ Club). Garden BirdWatch is organised by the BTO and funded by volunteers' subscriptions and donations. The London Bird Project was organised by the BTO and funded by the Bridge House Estate Trust.

37