Adapting conservation efforts to face climate change: Modifying nest-site provisioning for lesser kestrels

Biological Conservation 144 (2011) 1111–1119

Contents lists available at ScienceDirect

Biological Conservation

journal homepage: www.elsevier .com/ locate /biocon

Adapting conservation efforts to face climate change: Modifying nest-siteprovisioning for lesser kestrels

Inês Catry a,⇑, Aldina M.A. Franco b, William J. Sutherland a

a Conservation Science Group, Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, UKb School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, UK

a r t i c l e i n f o

Article history:Received 1 October 2010Received in revised form 21 December 2010Accepted 27 December 2010Available online 22 January 2011

Keywords:Climate changeAdaptationFalco naumanniLesser kestrelTemperatureArtificial nests

0006-3207/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.biocon.2010.12.030

⇑ Corresponding author. Tel.: +351 933193712.E-mail addresses: [email protected] (I. Catry)

Franco), [email protected] (W.J. Sutherland

a b s t r a c t

Adaptation to climate change has recently become a crucial element on the climate change policy agendaas it is now recognized that even the most stringent mitigation efforts may not arrest the effects of cli-mate warming. The ecological impacts and costs of predicted weather-related extreme events, such asextreme temperatures, are not fully understood and may present unexpected challenges to conservation-ists that require solutions. In Portugal, provisioning of artificial nests has been the main driver of thespectacular increase in the endangered lesser kestrel population. Nevertheless, atypically high tempera-tures recorded during the 2009 breeding season coincided with a mortality of 22% of surveyed chicks inprovided nests. Hot days did not affected prey delivery rates to the nestlings, suggesting that the die-offwas due to chicks’ acute dehydration. Chick mortality was significantly higher amongst younger individ-uals. Within survivors, physiological costs of high temperatures significantly affected chick growth andbody condition at fledging. Nest-site microclimate was influenced by nest-type and compass orientation:wooden nest-boxes attained the highest temperatures, exceeding 55 �C when facing south, so explainingthe recorded higher mortality, lower growth rates and lower fledging body condition among broods inthese nests. Simulated scenarios of global warming with increasing occupation rate of artificial nestsdue to reductions in alternatives predicted a reduction in population growth rate. In the worst scenario,with a 100% occupancy of nest-boxes, the population growth would decline on average 7% per year. Theimpact of high temperatures on lesser kestrel breeding success highlights a need for actions to modifyand research to adapt conservation efforts and future planning to account for climate change.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

The two main policy responses to global warming are mitiga-tion and adaptation (IPCC, 2007). While most of the debate hasbeen about mitigation to reduce the rate of climate change, thereis growing evidence that adaptation – ‘‘adjustment in natural orhuman systems in response to actual or expected climatic stimulior their effects’’ (IPCC, 2007) – will be essential to address its im-pacts given the unavoidable climate warming (IPCC, 2007; Pielkeet al., 2007; EEA, 2008). Research and policy on planning for adap-tation is emerging in some countries, states or provinces, but it isrecognized that most of the recommendations are too general tobe much use in practice (Galatowitsch et al., 2009; Heller andZavaleta, 2009).

Among the effects of global warming, extreme weather events(such as droughts, tropical cyclones, extreme temperatures andthe frequency and intensity of precipitation) are likely to have

ll rights reserved.

, [email protected] (A.M.A.).

the greatest negative impacts on natural systems (IPCC, 2007)and are predicted to increase in both frequency and magnitude(Easterling et al., 2000; Meehl and Tebaldi, 2004). Catastrophicmortality due to extreme weather events has occasionally beendocumented (Finlayson, 1932; Tompa, 1971; Garel et al., 2004;Welbergen et al., 2008), but few studies examined the impactson animal populations due to its rarity and unpredictability(Altwegg et al., 2006; Thibault and Brown, 2008). As anticipationof changes improves the capacity to adapt and reduce the risksof climate change, monitoring population responses to extremeevents is crucial for either planning or implementing conservationmeasures (IPCC, 2007; Rebetez et al., 2008). Moreover, recommen-dations for effective adaptation actions based on scientific researchare imperative to decision-makers and are likely to be of highestconcern for endangered species and regions where climate changeis projected to worsen conditions.

In this study, we examined the impacts of atypically high tem-peratures on the growth and survival of nestling lesser kestrelsFalco naumanni in the Mediterranean region; climate warmingis predicted to have particularly profound ecological effects in thisregion (Sala et al., 2000).

1112 I. Catry et al. / Biological Conservation 144 (2011) 1111–1119

Among birds, extreme air temperatures that exceed species’physiological limits can lead to large direct mortality due to dehy-dration (McKechnie and Wolf, 2010), especially among embryosand nestlings, which are more sensitive to thermal and hydric con-ditions at the nest (Lloyd and Martin, 2004).

The lesser kestrel is a small colonial falcon classified as globallythreatened due to the recent sharp declines in its WesternEuropean populations (BirdLife International, 2004). It is closelyassociated with open agricultural habitat with nesting coloniesusually either in isolated abandoned farmhouses or in buildingssuch as castles and churches, within villages or towns. Nests areusually located in cavities in the walls or under tiled roofs. InPortugal, the restoration of historic buildings and poaching withinurban areas has led to the extinction of most known urban colonies(Araújo, 1990). As a consequence, most of the population is now inrural areas, but the disintegration and collapse of farm buildingshas led to a shortage of nest-sites (Franco et al., 2005).

Conservation efforts, including a massive provisioning of artifi-cial nest-sites, has been shown to be an effective measure bymitigating the lack of traditional breeding sites, allowing the estab-lishment of new sites, by reducing predation rate and reducinginterspecific competition; as a result both breeding success andpopulation size have increased (Catry et al., 2009). In the future,as existing farm building continue to deteriorate, the lesser kestrelis likely to become closely associated with artificial breeding sitesand strongly dependent on the availability of conservation projectsfunds for their persistence (Catry et al., 2009). Therefore, it isimportant to determine the suitability of provided nest-sites bothfor present conditions and the most likely future scenarios in orderto ensure the species survival. While both the location and charac-teristics of artificial nest-sites can be selected to reduce the im-pacts of both predation and interspecific competition (Catryet al., 2009), there is no assessment of nest microclimate effectson lesser kestrel offspring.

During the 2009 breeding season, high chick mortality andweight loss was detected in artificial nest-sites, coincident withregistered extreme hot days. Evaluating the trade-offs betweennest-site characteristics and lesser kestrel breeding success underhigh temperature events can thus help predict the effects of suchevents in a future scenario of increasing temperatures and provideguidelines for adaptation actions. We focus on nestling growth be-cause it is an important component of fitness that is considerablyinfluenced by temperature through its effect on energy and waterbudgets (Gebhardt-Henrich and Richner, 1998; Visser, 1998). Thus,the aims of this study are: (1) relate chick growth and mortality todaily maximum temperatures, (2) compare chick growth and chickbody condition across various nest-types, (3) determine the rela-tionship between nest-type characteristics and location on nest-site microclimate, (4) evaluate the relationship between chickmortality and population demography and predict future trendsunder different scenarios of nest-site occupancy in a warmer worldand finally, and (5) provide recommendations for managementadaptation regarding nest-site provisioning for the species.

2. Methods

2.1. Study area

The study was conducted in the Castro Verde plains, Alentejo,Southern Portugal. This is the main area of cereal steppes inPortugal with a national and international importance regardingthe conservation of many endangered bird steppe species (Costaet al., 2003). The area is classified as a meso-mediterraneanbioclimatic stage (Rivas-Martínez, 1981), with hot dry summers(30–35 �C on average in July), fairly cold winters (averaging

5–8 �C in January) and over 75% of annual rainfall (500–600 mm)concentrated in October–March (Moreira et al., 2005). Presently,the Castro Verde Special Protection Area (SPA), comprising85,000 ha, contains 430 pairs (around 40 colonies), almost 80% ofthe total Portuguese lesser kestrel population (Catry et al., 2009).Most colonies are located either in old adobe-built abandonedfarmhouses (nests are located in cavities in walls or in the roof, un-der the tiles) or artificial nesting structures. Since 1996, but mostlyfrom 2003, about 800 artificial nest-sites were provided in the areaincluding new cavities opened in abandoned existing buildings,wooden nest-boxes, clay pots and breeding walls and towers(adobe and concrete-built structures with many available cavities).In 2007, about 50% (n = 212 pairs) of the Portuguese populationused artificial nests (Catry et al., 2009).

Lesser kestrels return to Castro Verde from their African winter-ing grounds mainly in early February (Catry et al., 2011) and laytypically 4–5 eggs in April and May (mean laying date = 28 Aprilfor 2003–2006, n = 1532; Henriques et al., 2006).

2.2. Data collection

Between 25 May and 5 July 2009, on a 2 day rotation, we visitedfour lesser kestrel colonies (range size: 19–66 pairs) to estimatenestling growth rates from hatching to fledging and assess chickmortality. Chicks were individually marked with cable ties athatching (and ringed later); body mass was weighed to the nearest0.1 g using a portable electronic balance. Monitored nests (all arti-ficial) included six stone cavities, six adobe cavities and four woo-den nest-boxes. Because, as described below, some chicks died onthe 12th of June, we chose six more wooden nest-boxes and threeconcrete cavities (in a breeding wall) to increase the sample, thesecontained chicks aged between 5 and 15 days. Mean brood sizewas 3.72 ± 1.1 for a total of 25 nests and 93 chicks.

2.2.1. Estimation of direct and indirect mortalityAmongst our sampled nests, 15 chicks were found dead across

the breeding period. Within a brood of six, two nestlings died fromstarvation and in other three broods, three chicks died from fatalinjuries. Amongst these five chicks, dead was preceded by a clearpattern of decrease in body weight. The remaining 10 chicks werefound dead inside the nests, with no sign of injuries, parasite infec-tion or predation. Nine died in the 12 of June (the first very hot day)and one in the hottest day. None of the deaths were preceded bylost of weight in previous days being therefore attributed to thehigh temperatures (direct mortality).

Within wooden nest-boxes, only four chicks disappeared duringthe entire chick rearing period, all coincident with the two hottestdays. Three chicks were found alive outside the nests, in a poorcondition (having lost 11, 20 and 26 g in 2 days); 2 days after beingreturned to their nests, chicks’ body condition improved 18, 27 and34 g, respectively. All missing chicks disappeared from woodennest-boxes with brood size lower than three (making less likelycompetition among siblings) and three of them belonged to nest-boxes where direct mortality occurred. Amongst adobe cavities,eight chicks disappeared during the study period, all in 2 days,after the first very hot day and during the hottest day. Missingchicks belonged to broods where all chicks lost significant weightduring the hottest days (see Section 3). Although some missingnestlings were slightly below the average brood weight, two ofthem had the best condition within their broods. Two missingchicks, older than 25 days, were excluded from mortality calcula-tions to preclude misclassifying early fledgling as mortality. Withinall monitored nests, we found no evidence of predation (predationamong monitored wooden nest-boxes was never detected in6 years) or ectoparasites. Thus, indirect mortality associated with

I. Catry et al. / Biological Conservation 144 (2011) 1111–1119 1113

high temperatures was attributed to four and six missing chicks inwooden nest-boxes and adobe cavities, respectively.

2.2.2. Effect of temperature on feeding ratesTo evaluate the effect of hot days on parental behaviour, which

could indirectly affect nestling growth (by reducing foragingcapacity and prey availability; Siikamäki, 1996), we recorded preydelivery rates (number of food items brought to the nest per hourand per chick) in twelve nests from 5 to 28 June 2009, with a totalof 24 h of observation per nest. Four observation periods werespread through the day (8:00–10:00, 10:00–12:00, 16:00–18:00and 18:00–20:00) avoiding the period of lower activity (deliveryrates within these periods were not significantly different,Kruskal–Wallis v2 = 5.9, p = 0.11). We then compared deliveryrates in days with maximum temperatures above and below37 �C. Data on daily maximum temperatures were obtained fromthe closest (about 40 km) meteorological station, in Beja.

2.2.3. Nest-site microclimateTo assess nest-site microclimate, forty-six micro-T DS1922L

temperature loggers (NexSens Technology, Fondriest Environment,USA) were deployed in five different artificial nest-types facing thenorth (cavities in adobe, concrete and stone walls, clay pots andwooden nest-boxes), recording daily temperatures at 60 min inter-vals between 5 June and 9 July 2010. Four devices were kept out-side (in shadow) to determine external air temperatures. Thedevices had an accuracy of ±0.5 �C, a resolution of 0.0625 �C andan operating range of �40 to 85 �C. From 12 to 24 of July, we testedthe effect of orientation on 31 wooden nest-boxes facing north,east, south or west and the difference between six clay pots thatwere painted white from six that were unpainted.

2.3. Statistical analyses

We used a generalized additive mixed model (GAMM) to de-scribe the relationship between chick growth and daily maximumtemperatures. For this paper, chick growth refers to the modelleddifference in chick mass between two consecutive visits (every2 days, at the same hour); thus being positive for mass gain or neg-ative for mass loss. Dead chicks were not weighed and were, there-fore, attributed with a mass loss of 25 g (mean loss of weightamong the lighter chicks during the hottest days). All chicks olderthan 25 days were excluded from the model, as slight variations inchicks’ weight close to fledgling date is often observed. The dailymaximum temperature was taken as the maximum temperatureregistered within each 2 days interval. We assumed a Gaussian er-ror distribution and an identity link function. The nestling age,brood size and nest-type were also included as predictors as theymay influence chick body mass; brood size was removed fromthe final model given its non-significant effect. The model was fit-ted using the ‘‘mgcv’’ package in R (R Development Core Team,2008) and the degree of smoothing for the non-linear terms (ageand maximum temperature) was estimated using cross-validation(Zuur et al., 2009). The use of mixed-effects models allowed us toinclude both chick and nest identity as random factors. Modelselection was based on identifying which explanatory variableshad significant effects. We checked that there were no trends inresiduals.

Nestling growth curves were estimated using a non-linearregression to fit a logistic growth curve for the entire data set ofnestlings weighed (n = 93). Three other growth curves were fitted,accounting for nestlings of each nest-type (nestlings from concretecavities were not included due to the small sample size and lack ofmass measurements in young chicks). Chick body condition atfledging was calculated using a body condition index BCI = 1 �(residual OM)/TM (Le Corre et al., 2003) where OM is the observed

mass (g), TM is the theoretical mass (g) calculated with the non-linear regression between mass and age, and residual OM is the dif-ference between TM and OM. With this equation, a BCI of 1 meansthat the individual has its normal mass for its age; BCI below 1means that the bird is lighter than expected (low body condition)and a BCI exceeding 1 means that the bird is heavier than expected(high body condition). Comparisons of chick body condition amongdifferent nest-types were assessed using a one-way ANOVA fol-lowed by post hoc Tukey tests.

We used a linear regression analysis to assess the relationshipbetween the amount of weight lost or gained in the first very hotday (12 June) and chick body condition at fledging. Differences be-tween delivery rates in days with maximum temperatures aboveand below 37 �C were compared by Mann–Whitney U tests.Nest-site microclimate amongst nest-types was compared usingone-way ANOVA and t-tests.

All analyses were performed with R (R Development Core Team,2008).

2.4. Demographic model and population trends

We used matrix population models to investigate potentiallong-term impacts of extreme temperatures on lesser kestrel pop-ulation dynamics through its effect on breeding success. Firstly, wedeveloped a female-based model to estimate the growth rate of thePortuguese lesser kestrel breeding population and then use thismodel to predict future trends under different simulated scenariosof nest-site occupancy and related chick mortality. Using programULM (Legendre and Colbert, 1995), we constructed a matrix popu-lation structured by population stages. The general formulation forthe matrix projection invariant in time takes the form Nt+1 = MNt,where M is the population projection matrix, incorporating dataon fertility and survival probabilities for each population stageand Nt is a vector with the abundance of individuals in each ofthe life-cycle stages (see Hiraldo et al. (1996) for details). The pop-ulation of females was divided in two age classes: yearlings andadults (individuals two or more years old). The demographicparameters used in the model were obtained from both our ownwork and literature. The number of adult and yearling femaleswas calculated based on the number of breeding pairs estimatedin 2007 (n = 540, Catry et al., 2009) and on the proportion of birdsthat attempt to breed (following the values of Hiraldo et al., 1996).Breeding success (2.006 ± 0.318) was assessed for the 2003–2007period, when most Portuguese colonies were monitored (LPN andI. Catry unpublished data). Adult survival probability was esti-mated from capture–recapture data of ringed lesser kestrels be-tween 1998 and 2005 in the two largest Portuguese colonies. Weused the Robson–Pollock model (Pollock, 1975), which is obtainedwhen fitting the Cormack–Jolly–Seber model separately for eachage class (Lebreton et al., 1992). Using program MARK (Whiteand Burnham, 1999) and starting from the full-time dependentmodel, we considered several sub-models, in which parameterdependence in relation to time was not included. We performedgoodness-of-fit tests using program U-CARE (Choquet et al.,2009) to test global homogeneity assumptions. The most parsimo-nious model was selected, using likelihood ratio tests and theAkaike Information Criterion (AIC). Estimation of survival (u) andrecaptures (p) probabilities was derived from maximum likelihoodmethods. According to the best fitting model, ua(�)pa(t), adult sur-vival was estimated at 0.71 ± 0.04 (for detailed methods andresults see Catry, 2005). For juvenile survival we used data fromthe French population (0.499 ± 0.021; Mihoub et al., 2010). Esti-mates of population growth rate and probability of extinction wereobtained by running 1000 Monte Carlo simulations during100 years. Stochastic models were used in order to incorporateenvironmental effects on demographic parameters (adult and

1114 I. Catry et al. / Biological Conservation 144 (2011) 1111–1119

juvenile survival, Legendre and Colbert, 1995). Based on the mainmodel, we estimated the population growth rate for three possiblescenarios by replacing the value of the breeding success in the pop-ulation projection matrix (Table 4). Breeding success was esti-mated by considering the breeding success used in the mainmodel corrected for chick mortality associated with the high tem-peratures assessed in 2009 for each artificial nest-type. Therefore,all the scenarios account for the same annual chick mortality(equal to 2009) being the number of breeding pairs in each nest-type that determines the final breeding success. In the first sce-nario, we considered that the proportion of breeding pairs amongstdifferent nests followed the distribution observed in 2007 for thewhole population (as the main model). The second and third sce-narios simulate the absence of lesser kestrel pairs breeding in nat-ural nest-sites. As in other countries, the Portuguese lesser kestrelpopulation is likely to become highly dependent on artificial nest-sites due to the high risk of collapse and restoration of abandonedfarmhouses (Catry et al., 2009). Thus, in the second scenario allpairs were distributed amongst provided artificial nest-types, inthe proportions observed in 2007. Since wooden nest-boxes areone of the most common nest-types provided worldwide for thisspecies (e.g. García and Guzmán, 1991; Bux et al., 2008; Schulmanet al., 2002; Ivanov, 2007), the third scenario considers only nest-boxes to be used by the entire population. For nest-types wherechick mortality due to high temperatures was not assessed (e.g.clay pots, nests under tiles) we took a conservative approach, con-sidering such chick mortality to be null.

Table 1Direct and indirect chick mortality expressed as absolute values and proportionsinflicted by high temperatures in each artificial nest-type. Sample sizes are given asboth chicks and nests.

Chick mortality Number ofchicks

Number ofnests

Direct Indirect

Stone cavities 0 (0.00) 0 (0.00) 26 6Adobe cavities 0 (0.00) 6 (0.23) 26 6Concrete cavities 0 (0.00) 0 (0.00) 10 3Wooden nest-boxes 10a (0.31) 4b (0.13) 32 10Total 10 (0.11) 10 (0.11) 93 25

a Two complete broods of three and four chicks and single chicks in two broods ofthree and one brood of two.

b Three of these chicks were prevented from certain death, after being rescuedfrom the ground in a weak condition.

3. Results

3.1. Daily maximum temperature, chick growth and mortality

Maximum temperatures exceeding 39 �C were registered for3 days during the study period (12, 20 and 21 of June), with anaverage maximum daily temperature of 36.9 �C (SE = 2.7) betweenthe first and last hottest day (Fig. 1). Twenty out of the 93 (22%)surveyed chicks died within this period. Both direct and indirectchick mortality coincident with high temperatures was observed:some chicks perished inside the nests (n = 10) while others jumpedfrom the nests and died from starvation, predation or heat (n = 10).The highest direct mortality occurred on the 12 June (maximumtemperature = 39.1 �C) after a relatively mild day (maximum =

Fig. 1. Daily maximum temperature (solid diamonds) in the study area from 25 May towere registered for 3 days. Average daily maximum temperatures (open circles) and sta

34 �C). During this day, chicks from three of seven nest-boxes notincluded in the study were also found dead. Direct mortality wasonly registered in wooden nest-boxes and affected younger chicks(11.6 ± 3.8 and 14.9 ± 4.4 days old for chicks that died and survivedin wooden nest-boxes, respectively; t = 2.13, df = 62, p = 0.037),whereas indirect mortality took place in both wooden nest-boxesand adobe cavities (Table 1).

Results from the GAMM model show a significant effect of dailymaximum temperature, chick age and nest-type on chick growth(Table 2). The fitted curves in Fig. 2 show that an increase in dailymaximum temperatures is associated with a higher probability oflosing weight. As expected, age was an important predictor of chickgrowth (Fig. 2d) and nest-type was also retained in the model,showing that chick growth in wooden nest-boxes is significantlylower than in other nest-types (Table 2, Fig. 2c). Overall, the threepredictors included in the model explained 61% of the variance inchick growth (Table 2).

Comparison of fitted growth curves for nestlings in differentnest-types is shown in Fig. 3a–c. Observed growth rates in bothadobe and stone cavities were coincident or higher than the aver-age growth rate (including all nestlings), whereas in wooden nest-boxes the opposite was observed. In adobe cavities, because chickswere of similar ages in all nests, the decline in chick growth asso-ciated with the hottest days is clearly visible (Fig. 3a). Nestlingsfrom wooden nest-boxes showed the highest residuals (|predicted

5 July 2009. Within this period, maximum temperatures above 39 �C (dashed line)ndard deviations from 2003 to 2008 are shown for comparison.

Table 2Generalised additive mixed model (GAMM) used to model lesser kestrel chick growthin relation to daily maximum temperature, age and nest-type. Estimation of thecoefficients (b) are given with standard errors (±SE); s(..) denotes predictors fitted asnon-parametric smoothing terms; edf is the estimated degrees of freedom, F-valuesand t-values are used to test the significance of non-parametric smoothing andparametric terms, respectively. Adjusted R2, scale estimate and sample size (n) valuesare shown.

Predictors b (± SE) edf F-value/t-value p-Value

s(max. temp) 3.9 95.93 <0.001s(max.temp): stone 3.47 21.94 <0.001s(max. temp): adobe 3.73 42.84 <0.001s(max. temp): nest-box 3.74 53.72 <0.001s(age) 6.8 45.35 <0.001

Nest-typeIntercept 10.68 ± 0.49 21.84 <0.001Wooden nest-box �1.75 ± 0.77 �2.26 0.025Stone cavity 1.46 ± 0.68 2.16 0.031

Adj. R2 = 0.61, scale est. = 61.57 n = 745

I. Catry et al. / Biological Conservation 144 (2011) 1111–1119 1115

mass minus observed mass|), with several chicks weighting muchless than predicted by the growth curve (Fig. 3b).

We found a significant positive relationship between theamount of mass gained (or lost) in the first of the hottest daysand chick body condition at fledging (R2 = 0.26, F1,38 = 14.86,

Fig. 2. Effect of maximum temperature in: (a) stone cavities, (b) adobe cavities and (c)from the generalized additive mixed model analysis (GAMM). The y-axis shows the contrvariable (chick growth). Ticks in the x-axis represent location of observations along the pshown.

p = 0.0004), showing that chicks do not recover totally from previ-ous losses of weight. Chicks’ body condition index was significantlydifferent among nest-types (F2,49 = 3.79, p = 0.029, followed by posthoc Tukey tests) showing significant differences between woodennest-boxes and stone cavities (means ± SE adobe cavities: 1.01 ±0.09, n = 15; wooden nest-boxes: 0.94 ± 0.13, n = 18; and stonecavities: 1.05 ± 0.09, n = 21; Fig. 3d).

3.2. Effect of temperature on feeding rates

Results from colony observations revealed no significant differ-ences on prey delivery rates during days with maximum tempera-ture below compared with those above 37 �C (1.62 ± 0.94 and1.59 ± 0.31 prey h�1 chick�1; mean temp = 28.0 ± 4.0 and 37.6 ±0.6 �C, respectively; Mann–Whitney U test W = 831, p = 0.65). Theage of chicks was similar in both sample periods (15.9 ± 7.9 and15.2 ± 4.8 days Mann–Whitney U test W = 907, p = 0.89).

3.3. Nest-site microclimate

Fig. 4a shows the maximum temperatures registered in eachnest-type for 8 days in 2010, when the exterior temperature ex-ceeded 35 �C. Mean daily temperature inside wooden nest-boxeswas significantly higher than in any other nest-type and clay pots

wooden nest-boxes and (d) age on lesser kestrel chick growth (solid line) obtainedibution of the fitted centred smooth terms s(names of the predictor) to the responseredictors. Partial residuals (points) and ±95% confidence intervals (dashed lines) are

0 5 10 15 20 25 30 35

0

20

40

60

80

100

120

140

160

180

200

chick age (days)

mas

s (g

)

adobe cavitiesall nests

0 5 10 15 20 25 30 35

0

20

40

60

80

100

120

140

160

180

200

chick age (days)

mas

s (g

)

wooden nestboxesall nests

0 5 10 15 20 25 30 35

0

20

40

60

80

100

120

140

160

180

200

chick age (days)

mas

s (g

)

stone cavitiesall nests

adobe cavity stone cavity wooden nestbox0.85

0.90

0.95

1.00

1.05

1.10

1.15

1.20

nest type

chic

k bo

dy c

ondi

tion

(a)

(c)

(b)

(d)

Fig. 3. Comparison of nestling growth curves (a–c) and chick body condition index at fledging (d) in adobe cavities, wooden nest-boxes and stone cavities. Logistic curveswere fitted separately for chicks in each nest-type (solid line) and all nestlings (dashed line). Solid diamonds correspond to observation values used to fit the curves; values of‘‘zero’’ for chick mass represent dead or missing chicks and were not included for growth curve estimation. The arrow in panel a shows the effect of the first very hot day (12June). In panel d, an index of 1 means that the individual has its normal mass according to its age; below 1 means that the bird is lighter than expected (low body condition)and a over 1 means that the bird is heavier than expected.

stone adobe concrete clay pot nest−box outside

30

35

40

45

50

nest−type

max

imum

tem

pera

ture

(ºC

)

north east west south outside

25

30

35

40

45

50

55

nest orientation

max

imum

tem

pera

ture

(ºC

)

(a) (b)

Fig. 4. Maximum daily temperatures recorded in: (a) different nest-types and (b) wooden nest-boxes facing different orientations during the chick rearing period in 2010. Airtemperature in the exterior (grey box) is shown for comparison. Median, upper and lower quartiles (box) and the range (whiskers) are represented.

1116 I. Catry et al. / Biological Conservation 144 (2011) 1111–1119

were significantly hotter than adobe and stone cavities (Table 3).During the hottest day (air temp = 40.0 ± 1.5 �C) the temperaturereached 47.2 ± 7.4 �C inside wooden nest-boxes (n = 6), 44.6 ±4.0 �C in clay pots (n = 7) and 41.1 ± 0.7 �C in cavities in concrete

walls (n = 6). Among cavities located in adobe or stone walls thedaily maximum temperatures were always lower than the outsidetemperature, reaching 37.9 ± 1.2 �C (n = 6) and 38.2 ± 0.7 �C (n = 7)during the hottest day, respectively (Fig. 4a). Temperature

Table 3Effect of nest-type (wooden nest-boxes, clay pots and stone, adobe and concrete cavities), nest orientation (wooden nest-boxes facing north, south, east andwest) and solar protection (white painted and unpainted clay pots) on nest-site microclimate (maximum temperatures) among provided nests for lesserkestrel in Castro Verde, Portugal.

Variable F-value/t-value p-Value Temperature effect

Nest-type F4, 35 = 18.0 <0.001 wooden nest-boxes > all other nests; clay pots > adobe and stoneOrientation F3, 48 = 12.9 <0.001 north < south, west and eastSolar protection t24 = 2.5 0.020 unpainted > white painted

Table 4Lesser kestrel population growth rate (k) estimated for the present and future possible scenarios according to differential nest-site occupancy and associated chickmortality due to high temperatures. In the main model and scenario 1, the proportion of pairs in each nest-type follows the distribution assessed in 2007 while inscenarios 2 and 3 all breeding pairs use artificial nest-sites (see Section 2). Chick mortality rate in adobe cavities and wooden nest-boxes is 0.23 and 0.44, respectively.All models assume a breeding population of 540 pairs. Breeding success represents the number of fledged chicks per pair.

Breeding success Population growth (k) Chick mortality Proportion of breeding pairs

Natural nests Artificial nests

Adobe Nest-box Other

Main model 2.0 1.08 ± 0.0003 No 0.48 0.28 0.09 0.15Scenario 1 1.8 1.05 ± 0.0002 Yes 0.48 0.28 0.09 0.15Scenario 2 1.6 1.01 ± 0.0002 Yes 0.00 0.54 0.17 0.30Scenario 3 1.12 0.93 ± 0.0001 Yes 0.00 0.00 1.00 0.00

I. Catry et al. / Biological Conservation 144 (2011) 1111–1119 1117

variability between wooden nest-boxes attached to walls with dif-ferent orientations is showed in Fig. 4b. The north-oriented nest-boxes had significantly lower daily maximum temperatures(33.5 ± 4.1) than the west (38.3 ± 4.0), east (42.1 ± 3.6) and south-oriented ones (42.9 ± 5.3; Table 3); the last ones reached the high-est temperatures, achieving 54.8 ± 2.4 �C (n = 8) with a38.5 ± 1.6 �C (n = 4) outside temperature (Fig. 4b). Finally, dailymaximum temperatures were significantly higher in unpaintedclay pots (43.7 ± 0.6 �C, n = 6) compared to white painted ones(40.0 ± 3.4 �C, n = 4, Table 3).

3.4. Effect of chick mortality on population dynamics

The main model, not accounting for chick mortality due to hightemperatures, predicted a long-term growth rate of 1.08(SE = 0.0003), with an average population increase of 8% per year(Table 4). However, when considering an annual constant mortal-ity (equal to the one reported in 2009, scenario 1), the populationgrowth rate decreases to 1.05 (SE = 0.0002). The two scenarios thatassume all lesser kestrel pairs breeding in artificial nests wouldreduce even more the population growth rate (Table 4). Whileunder scenario 2 (most pairs using artificial adobe cavities) thepopulation would remain stable (k = 1.01 ± 0.0002): under scenario3 (all pairs using wooden nest-boxes) the population woulddecline on average 7% per year (k = 0.93 ± 0.0001, Table 4). Underscenario 3, the probability of extinction would be 99% in 110 years,although with such a decline density dependence is likely toslightly reduce the extinction rate. We should stress that ourapproaches are conservative, considering no mortality in concretecavities or clay pots (where maximum temperatures exceed theones in adobe cavities).

4. Discussion

Rapid global warming will continue for several decades what-ever future actions are taken, causing inevitable impacts acrossspecies and ecosystems (e.g., Sæther et al., 2000; Root et al.,2003; Crick, 2004; Thomas et al., 2004; Huntley et al., 2007; Amanoet al., 2010). Nevertheless, there has been little consideration of thepossible adaptation strategies for maintaining biodiversity (Hellerand Zavaleta, 2009; Senapathi, 2010) and even fewer examples oftesting their efficiency (but see Willis et al., 2009). Increasing con-

nectivity to allow species to move significant distances, promotelandscape level heterogeneity to enhance species to stay withintheir climate envelope and assisted colonisation or translocationhave been referred as the most favourable options to face climatechange (Hopkins et al., 2007; Heller and Zavaleta, 2009; Hodgsonet al., 2009; Willis et al., 2009; Sutherland et al., 2010). In the otherhand, the potential to adaptive management for individual sitesand populations to increase their resilience and track climatechange has received little attention (but see Pearce-Higgins,2010). Overall, the absence of climate change predictions for spe-cific areas, uncertainty about the species capacity to adapt and lackof evidence that the proposed actions will be effective are delayingthe implementation of specific adaptation plans (Galatowitschet al., 2009).

This study provides an example of how climate change maypresent unexpected challenges to conservationists and how con-servation may adapt to minimize its impacts. In 2009, the occur-rence of unusual high temperatures during the lesser kestrelbreeding season provided us with a unique opportunity to evaluatethe effects of such extreme weather events on chick growth andmortality, predict the impacts of its increasing frequency on popu-lation dynamics and recommend effective adaptation actions.

4.1. Effect of high temperatures on chicks’ fitness and survival

During the 2009 breeding season 22% of the surveyed chicksdied, coinciding with registered high temperatures. Comparisonof prey delivery rates between hot and mild days was similar, sug-gesting that reductions in offspring number and weight (to up 30 gor 27% of body mass) reflected the direct increasing costs of ther-moregulation (Bradley et al., 1997; Lloyd and Martin, 2004), ratherthan a decrease in the availability and/or accessibility of preys. In-side the nests and during extremely hot days, thermoregulatorywater requirements increase and at environmental temperaturesexceeding the normal body temperature, chicks overheat quicklybecause of their relatively large surface area and low heat capacity(Elkins, 1983; Visser, 1998). Because mass-specific evaporativewater loss rates increases with decreasing body mass, vulnerabilityto acute dehydration is most pronounced in smaller individuals(McKechnie and Wolf, 2010), explaining the higher mortality re-ported among younger nestlings. For the same reason, the abnor-mal mortality may have been caused essentially by the hottest

1118 I. Catry et al. / Biological Conservation 144 (2011) 1111–1119

days being early in the season, affecting mostly young chicks. Inour study area, temperatures above 39 �C are fairly common duringthe summer but daily maximum temperatures of 39 �C in the firsthalf of June have been registered only in three amongst the last37 years (http://www.ncdc.noaa.gov/oa/ncdc.htlm). Mass loss dur-ing the chick rearing period significantly affected chick body con-dition at fledging. As chicks that fledge in poor condition mayhave reduced fat reserves and are less likely to survive throughoutmigration, chick body condition at fledging can be a good indicatorof survival probability (Rodríguez and Bustamante, 2003) and thus,high temperatures during chick rearing can have a medium to longterm negative impact on juvenile survival.

4.2. Effect of nest-type on nest microclimate and offspring fitness

Nest-site characteristics and orientation can have a strong im-pact on nest-site selection and reproductive success through its ef-fects on microclimate (e.g. Valkama and Korpimäki, 1999; Wiebe,2001; Lloyd and Martin, 2004; Hilton et al., 2004; Butler et al.,2009). Our results show that during hot days, nest temperaturein clay pots but especially in wooden nest-boxes can reach veryhigh values, several degrees above the air temperature. Chick mor-tality was reported in wooden nest-boxes facing south, west andeast but not in nest-boxes facing north, where the temperaturewas significantly cooler than in the other orientations. Despitebeing large (50 � 25 � 25 cm), wooden nest-boxes have no ventila-tion and, when exposed to the sun, interior temperatures can ex-ceed 55 �C. The highest temperatures reported inside woodennest-boxes can explain the observed pattern of nestling growthcurves, the lowest value of chick body condition and the highestmortality rate between monitored nest-types. Despite not havingbeen monitored for chick growth and survival, the high tempera-tures achieved in clay pots here (up to 48 �C) and in previous stud-ies (49 and 55 �C, Pomarol, 1996 and Tella et al., 1994, respectively)suggests that some chick mortality must occur during the hottestdays, particularly in nests facing south and unpainted ones. Adobeand stone cavities were the cooler nests, with mean daily maxi-mum temperatures slightly below the air temperature. Both mate-rials were traditionally used in house building in the region andmost natural nests are located in these structures. The observeddifference in chick mortality observed between these nests(0% and 23% indirect mortality in stone and adobe cavities, respec-tively) may be explained by its orientation, dimension or numberof fledglings inside the nests. Contrarily to the monitored adobecavities, all stone cavities faced the north, probably contributingto a more suitable nest microclimate, as suggested by the lowerlosses of weight registered within these nests. The size of the innerchamber and number of chicks may also affect nest microclimate.Among adobe cavities, unoccupied nests were significant coolerthan occupied ones (around 5 �C, IC unpublished data) and in smal-ler cavities evaporative water loss and dehydration may forcefledglings out of nests.

4.3. Effect of chick mortality on population dynamics and managementimplications

In southern Europe, climate change is projected to worsen con-ditions in a region already vulnerable to climate variability (Meehland Tebaldi, 2004; IPCC, 2007). Future climatic scenarios for Portu-gal project increases in summer average maximum temperaturesof 4–4.5 �C for the 2040s, and a significant increase in the occur-rence and duration of heat waves (Santos et al., 2001; Casimiroet al., 2006). In a warmer world, chick mortality in provided artifi-cial nests, along with the likely increase of their occupancy, is pre-dicted to preclude the recent positive population growth (scenario2) and even lead to the species extinction, as predicted if all pairs

breed in wooden nest-boxes (scenario 3). Moreover, the expecteddecrease in the juvenile survival associated with high tempera-tures could significantly increase the negative impact on popula-tion growth as population dynamics in lesser kestrels are mostsensitive to adult and juvenile survival than to fecundity (Hiraldoet al., 1996).

Wooden nest-boxes and clay pots are easy to install in anybuilding (new or old) and have been provided worldwide. Never-theless, the high temperatures inside these nests may result in eco-logical traps for the species and its future provisioning should beavoided. To reduce inside temperatures, re-designed woodennest-boxes could be installed under roof tiles (Pomarol, 1996)and other materials should be considered (as woodcrete; http://oryxdistribuciones.com). For those nests already provided, newdeployment should be considered in walls not exposed (or at sha-dow) and clay pots should regularly be white painted (see alsoPomarol, 1996). Adobe, stone and concrete are the most suitablematerials and should be used whenever is possible when providingnew breeding structures. Partial reopening of cavities in buildingsinstead of nest-box installation, non-exposed locations and largeinner chambers accounting for broods of five chicks would helpmaintain tolerable temperatures.

As in other species of conservation concern, nest-site provision-ing is considered to be a key factor for the recover of the Portugueselesser kestrel population in the last decade and suggested to becrucial for its maintenance in the near future (Catry et al., 2009).However, rapid global warming may jeopardize the effectivenessof present and future conservation strategies. This study enhancesthe importance of addressing the effects of climate change at localor species-level in order to help adjusting conservation actions toincrease their resilience to climate change in a more effectiveway. Conservation planning and climate adaptation strategiesshould be considered in the implementation of future conservationprojects and integrated in the elaboration and review of the Spe-cies Actions Plans.

Acknowledgements

Special thanks to T. Catry and M. Gomes for the valuable helpduring field work. LPN and P. Rocha provided important data onbreeding success from 2003 to 2006 and marked recaptured birds.Temperature dataloggers were funded by the NERC Ecology andHydrology Funding Initiative, NE/F009836/1 and A. Suggitt pro-vided valuable advice on their use. We are grateful to T. Amanofor the statistical help with the GAMM analysis and three anony-mous referees for useful comments to the manuscript. I.C. wasfunded by a Portuguese doctoral Grant from Fundação para aCiência e Tecnologia (SFRH/BD/28023/2006) and W.J.S. by theArcadia Fund.

References

Altwegg, R., Roulin, A., Kestenholz, M., Jenni, L., 2006. Demographic effects ofextreme winter weather in the barn owl. Oecologia 149, 44–51.

Amano, T., Richard J. Smithers, R.J. Tim H. Sparks, T.H., Sutherland, W.J., 2010. A250-year index of first flowering dates and its response to temperature changes.Proceedings of the Royal Society B. doi:10.1098/rspb.2010.0291.

Araújo, A., 1990. Dados sobre a evolução das populações de Peneireiro-das-torresFalco naumanni em Portugal, com especial incidência sobre o período de 1975 a1990. Serviço Nacional de Parques, Reservas e Conservação da Natureza (ICN),Lisboa.

BirdLife International, 2004. Birds in Europe: Population Estimates, Trends andConservation Status. BirdLife International, Cambridge, UK (Conservation SeriesNo. 12).

Bradley, M., Johnstone, R., Court, G., Duncan, T., 1997. Influence of weather onbreeding success of peregrine falcons in the Arctic. The Auk 114, 786–791.

Butler, M.W., Whitman, B.A., Dufty Jr., A.M., 2009. Nest box temperature andhatching success of the American kestrels varies with nest box orientation.Wilson Journal of Ornithology 121, 778–782.

I. Catry et al. / Biological Conservation 144 (2011) 1111–1119 1119

Bux, M., Giuseppe, G., Gustin, M., 2008. Nest box provision for lesser kestrel Falconaumanni populations in the Apulia region of southern Italy. ConservationEvidence 5, 58–61.

Casimiro, E., Calheiros, J., Santos, F.D., Kovats, S., 2006. National assessment ofhuman health effects of climate change in Portugal: approach and key findings.Environmental Health Perspectives 114, 1950–1956.

Catry, I., 2005. Análise da sobrevivência do Peneireiro-das-torres (Falco naumanni)com base em dados de captura–recaptura. M.Sc. Thesis. Instituto Superior deAgronomia, Universidade Técnica de Lisboa, Portugal.

Catry, I., Alcazar, R., Franco, A.M.A., Sutherland, W.J., 2009. Identifying theeffectiveness and constraints of conservation interventions: a case study ofthe endangered lesser kestrel. Biological Conservation 142, 2782–2791.

Catry, I., Dias, M.P., Catry, T., Afanasyev, V., Fox, J., Franco, A.M.A., Sutherland, W.J.,2011. Individual variation in migratory movements and winter behaviour ofIberian Lesser Kestrels Falco naumanni revealed by geolocators. Ibis 153, 154–164.

Choquet, R., Lebreton, J.D., Gimenez, O., Reboulet, A.M., Pradel, R., 2009. U-CARE:utilities for performing goodness of fit tests and manipulating CApture–REcapture data. Ecography 32, 1071–1074.

Costa, L.T., Nunes, M., Geraldes, P., Costa, H., 2003. Zonas Importantes para as Avesem Portugal. Sociedade Portuguesa para o Estudo das Aves, Lisboa.

Crick, H.Q.P., 2004. The impact of climate change on birds. Ibis 146, 48–56.EEA, 2008. Impacts of Europe’s Changing Climate – 2008 Indicator-Based

Assessment (No. 4/2008). European Environmental Agency, Copenhagen.Easterling, D.R., Meehl, G.A., Parmesan, C., Changnon, S.A., Karl, T.R., Mearns, L.O.,

2000. Climate extremes: observations, modeling and impacts. Science 289,2068–2074.

Elkins, N., 1983. Weather and Bird Behaviour. T & A.D. Poyser, London.Finlayson, H.H., 1932. Heat in the interior of South Australia and in Central

Australia: holocaust of bird-life. South Australian Ornithologist 11, 158–160.Franco, A.M.A., Marques, J.T., Sutherland, W.J., 2005. Is nest-site availability limiting

lesser kestrel populations? A multiple scale approach. Ibis 147, 657–666.Galatowitsch, S., Frelich, L., Phillips-Mao, L., 2009. Regional climate change

adaptation strategies for biodiversity conservation in a midcontinental regionof North America. Biological Conservation 142, 2012–2022.

García, A.S., Guzmán, J.M.S., 1991. Resultados de ocupacion de cajas anidaderas entendidos electricos en Extremadura (oeste de España): 1986–1990. Ecología 5,375–381.

Garel, M., Loison, A., Gaillard, J.M., Cugnasse, J.M., Maillard, D., 2004. The effects of asevere drought on mouflon lamb survival. Proceedings of the Royal Society B271, S471–S473.

Gebhardt-Henrich, S., Richner, H., 1998. Causes of growth variation and itsconsequences for fitness. In: Starck, J.M., Ricklefs, R.E. (Eds.), Avian Growthand Development: Evolution within the Altricial–Precocial Spectrum. OxfordUniversity Press, Oxford, pp. 324–339.

Heller, N.E., Zavaleta, E.S., 2009. Biodiversity management in the face of climatechange: a review of 22 years of recommendations. Biological Conservation 142,14–32.

Henriques, I., Constantino, R., Alcazar, R., 2006. Monitorização das colónias dePeneireiro-das-torres Falco naumanni em Portugal. LIFE Project LIFE 2002/NAT/P8481. LPN, Lisboa.

Hilton, G.M., Hansell, M.H., Ruxton, G.D., Reid, J.M., Monaghan, P., 2004. Usingartificial nests to test the importance of nesting material and nest shelter forincubation energetics. Auk 121, 777–787.

Hiraldo, F., Negro, J.J., Donázar, J.A., Gaona, P., 1996. A demographic model for apopulation of the endangered lesser kestrel in southern Spain. Journal ofApplied Ecology 33, 1085–1093.

Hodgson, J.A., Thomas, C.D., Wintle, B.A., Moilanen, A., 2009. Climate change,connectivity and conservation decision making: back to basics. Journal ofApplied Ecology 46, 964–969.

Hopkins, J.J., Allison, H.M., Wasmsley, C.A., Gaywood, M., Thurgate, G., 2007.Conserving Biodiversity in a Changing Climate: Guidance on Building Capacityto Adapt. DEFRA, London.

Huntley, B., Green, R.E., Collingham, Y.C., Willis, S.G., 2007. A Climatic Atlas ofEuropean Breeding Birds. Lynx Edicions, Barcelona.

IPCC, 2007. Climate Change 2007: Fourth Assessment Report of theIntergovernmental Panel on Climate Change. Cambridge University Press,Cambridge, UK.

Ivanov, H., 2007. Adapting Human Facilities to Safeguard the Threatened BirdSpecies in their Modern Homes in Macedonia. Final Report to Rufford SmallGrants, 2004–2007. FWFF, Macedonia.

Le Corre, M., Cherel, Y., Lagarde, F., Lormée, H., Jouventin, P., 2003. Seasonal andinter-annual variation in the feeding ecology of a tropical oceanic seabird, thered-tailed tropicbird Phaeton rubricauda. Marine Ecology Progress Series 255,289–301.

Lebreton, J.D., Burnham, K.P., Clobert, J., Anderson, D.R., 1992. Modeling survival andtesting biological hypotheses using marked animals: a unified approach withcase studies. Ecological Monographs 62, 67–118.

Legendre, S., Colbert, J., 1995. ULM, a software for conservation and evolutionarybiologists. Journal of Applied Statistics 22, 817–834.

Lloyd, J.D., Martin, T.E., 2004. Nest-site preference and maternal effects on offspringgrowth. Behavioral Ecology 15, 816–823.

McKechnie, A.E., Wolf, B.O., 2010. Climate change increases the likelihood ofcatastrophic avian mortality events during extreme heat waves. Biology Letters6, 253–256.

Meehl, G.A., Tebaldi, C., 2004. More intense, more frequent and longer lasting heatwaves in the 21st century. Science 305, 994–997.

Mihoub, J.B., Gimenez, O., Pilard, P., Sarrazin, F., 2010. Challeging conservationof migratory species: Sahelian rainfalls drive first-year survival of thevulnerable lesser kestrel Falco naumanni. Biological Conservation 143, 839–847.

Moreira, F., Beja, P., Morgado, R., Reino, L., Gordinho, L., Delgado, A., Borralho, R.,2005. Effects of field management and landscape context on grasslandwintering birds in Southern Portugal. Agriculture, Ecosystems andEnvironment 109, 59–74.

Pearce-Higgins, J.W., 2010. A worked example of climate change adaptation forsouthern range margin populations. BOU Proceedings – Climate Change andBirds. <http://www.bou.org.uk/bouprocnet/ccb/pearce-higgins.pdf>.

Pielke, R., Prins, G., Rayner, S., Sarewitz, D., 2007. Lifting the taboo on adaptation.Nature 445, 597–598.

Pollock, K.H., 1975. A K-sample tag-recapture model allowing for unequal survivaland catchability. Biometrika 62, 577–583.

Pomarol, M., 1996. Artificial nest structure design and management implications forthe lesser kestrel. Journal of Raptor Research 30, 169–172.

R Development Core Team, 2008. R: a language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna, Austria. <http://www.Rproject.org> (accessed 20.03.10).

Rebetez, M., Dupont, O., Giroud, M., 2008. An analysis of the July 2006 heatwaveextent in Europe compared to the record year of 2003. Theoretical and AppliedClimatology 95, 1–7.

Rivas-Martínez, S., 1981. Les étages bioclimatiques de la végétation de la PeninsuleIbérique. Anales del Jardín Botánico de Madrid 37, 251–268.

Rodríguez, C., Bustamante, J., 2003. The effect of weather on breeding success: canclimate change explain historical population declines? Journal of AnimalEcology 72, 793–810.

Root, T.L., Price, J.T., Hall, K.R., Schneider, S.H., Rosenzweig, C., Pounds, J.A., 2003.Fingerprints of global warming on wild animals and plants. Nature 421, 57–60.

Sæther, B.E.J., Tufto, S.E., Engen, S., Jerstad, K., Røstad, O.W., Skåtan, J.E., 2000.Population dynamical consequences of climate change for a small temperatesongbird. Science 287, 854–856.

Sala, O.E., Chapin III, F.S., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., Huber-Sanwald, E., Huenneke, L.F., Jackson, R.B., Kinzing, A., Leemans, R., Lodge, D.M.,Mooney, H.A., Oesterheld, M., Poff, N.L., Sykes, M.T., Walker, B.H., Walker, M.,Wall, D.H., 2000. Global diversity scenarios for the year 2100. Science 287,1770–1774.

Santos, F.D., Forbes, K., Moita, R. (Eds.), 2001. Climate Change in Portugal. Scenarios,Impacts and Adaptation Measures. Executive Summary and Conclusions. SIAM,Gradiva, Lisbon.

Schulman, I., Alon, D., Frankel, M., Leshem, Y., 2002. Conservation of the LesserKestrel in Israel. <http://www.birds.org.il/sip_storage/files/1/761.pdf>(accessed 20.08.10).

Senapathi, D., 2010. Climate change and birds: adaptation, mitigation & impacts onAvian populations. A Report on the BOU’s Annual Conference Held at theUniversity of Leicester, 6–8 April 2010. Ibis 152, 869–872.

Siikamäki, P., 1996. Nestling growth and mortality of Pied Flycatchers Ficedulahypoleuca in relation to weather and breeding effort. Ibis 138, 471–478.

Sutherland, W.J., Clout, M., Côté, I.M., Daszac, P., Depledge, M.H., Fellman, L.,Fleishman, E., Garthwaite, R., Gibbons, D.W., De Lurio, J., Impey, A.J., Lickorish, F.,Lindenmayer, D., Madgwick, J., Margerison, C., Maynard, T., Peck, L.S., Pretty, J.,Prior, S., Redford, K., Scharlemann, J.P.W., Spalding, M., Watkinson, A., 2010. Ahorizon scan of global conservation issues for 2010. Trends in Ecology andEvolution 25, 1–7.

Tella, J.L., Sánchez, I., Hiraldo, F., Donázar, J.A., 1994. Evaluación de nidalesartificiales para el cernícalo primilla. Quercus 97, 4–6.

Thibault, K.M., Brown, J.H., 2008. Impact of an extreme climatic event oncommunity assembly. Proceedings of the National Academy of Sciences 105,3410–3415.

Thomas, C.D., Cameron, A., Green, R.E., Bakkenes, M., Beaumont, L.J., Collingham,Y.C., Erasmus, B.F.N., Siqueira, M.F., Grainger, A., Hannah, L., Hughes, L., Huntley,B., van Jaarsveld, A.S., Midgley, G.F., Miles, L., Ortega-Huerta, M.A., Peterson, A.T.,Phillips, O.L., Williams, S.E., 2004. Extinction risk from climate change. Nature427, 145–214.

Tompa, F.S., 1971. Catastrophic mortality and its population consequences. Auk 88,753–759.

Valkama, J., Korpimäki, E., 1999. Nestbox characteristics, habitat quality andreproductive success of Eurasian kestrels. Bird Study 46, 81–88.

Visser, G.H., 1998. Development of temperature regulation. In: Starck, J.M., Ricklefs,R.E. (Eds.), Avian Growth and Development: Evolution within the Altricial–Precocial Spectrum. Oxford University Press, Oxford.

Welbergen, J.A., Klose, S.M., Markus, N., Eby, P., 2008. Climate change and the effectsof temperature extremes on Australian flying-foxes. Proceedings of the RoyalSociety B 275, 419–425.

White, C.G., Burnham, K.P., 1999. Program MARK: survival estimation frompopulations of marked animals. Bird Study 46, 120–138.

Wiebe, K.L., 2001. Microclimate of tree cavity nests: is it important for reproductivesuccess in Northern flickers? Auk 118, 414–421.

Willis, S.G., Hill, J.K., Thomas, C.D., Roy, D.B., Fox, R., Blakeley, D.S., Huntley, B., 2009.Conservation Letters 2, 45–51.

Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., Smith, G., 2009. Mixed EffectsModels and Extensions in Ecology with R. Springer, New York.