Chambers-L.E._2006 - La influencia de la variabilidad climática sobre los número de tres especies...

7/31/2019 Chambers-L.E._2006 - La influencia de la variabilidad climtica sobre los nmero de tres especies de aves acutic

1/13

Int J Biometeorol (2006) 50: 292 304DOI 10.1007/s00484-005-0019-2

ORIGINAL ARTICLE

Lynda E. Chambers . Richard H. Loyn

The influence of climate variability on numbers of three waterbirdspecies in Western Port, Victoria, 1973 2002

Received: 1 April 2005 / Revised: 9 November 2005 / Accepted: 14 November 2005 / Published online: 25 January 2006# ISB 2006

Abstract Seasonal and annual movements of Australianwaterbirds are generally more complex than those of their Northern Hemisphere counterparts, and long-term data are

needed to understand their relationships with climaticvariables. This paper explores a long-term (1973 2002)set of waterbird counts from coastal Victoria and relatesthem to climatic data at local and continental scales. Threespecies (Black Swan Cygnus atratus , White-faced Heron Egretta novaehollandiae and Grey Teal Anas gracilis ) werechosen for this analysis. Black Swans have large local breeding populations near the study region; White-facedHerons have smaller local breeding populations and GreyTeal breed extensively in ephemeral inland floodplains, suchas those in the Murray-Darling Basin. All showed significant relationships with streamflow, regional rainfall and theSouthern Oscillation Index (SOI) at appropriate scales and

time-lags, with streamflow explaining the most variance.Black Swans showed a strong seasonal cycle in abundanceand local climate variables had the greatest influence on thecounts. Numbers were positively correlated with streamflowin southern Victoria three to six seasons before each count.Broader-scale climatic patterns were more important for theother two species. Numbers of White-faced Herons were positively correlated with streamflow or rainfallover various parts of Australia seven to nine seasons before each count. Numbers of Grey Teal showed weak seasonal cycles, andwere negatively correlated with rainfall in Victoria or theMurray-Darling Basin in the seasons before or during eachcount, and positively with streamflow in the Murray-Darling

Basin 15

18 months before each count.

Keywords Australia . Climate . Rainfall . Streamflow .Waterbirds

Introduction

Climatic variations can have profound effects on ecologicaland human environments, and many of the ecological effectsremain poorly understood despite enormous advances inmeteorology (Nicholls 2001 ; Stevermer 2001 ; Pearce 2002 ).Seasonal and annual movementsof Australian waterbirds aregenerally more complex and less predictable than thoseof their Northern Hemisphere counterparts (Frith 1982 ;Kingsford and Norman 2002 ). This is partly related to the physiography and erratic climate of Australia (Pittock 2003 ;Wright 2004 ) and, in particular, to the highly variable

wetland and floodplain systems. Much of Australia experi-ences an arid climate, and wetlands may be dry for manyyears before filling with flushes of water originating fromrainfall events hundreds of kilometres away. These eventscan occur at any time of year, as a consequence of weather patterns in the tropical north (where most rain falls insummer) or the temperate south (where more rain falls inwinter and spring, and snow melts from the mountainsin spring). For waterbirds, population boom periods can beinitiated by such flooding, and bust periods can followmonths or years of drought. Even in a dry year, ephemeralfloodwaters from distant rains may allow arid parts of Australia to support many millions of waterbirds (Roshier et

al. 2001a , b). In the Northern Hemisphere, most waterfowlundertake regular and predictable migrations, in which most of the population moves between breeding areas andwintering areas each year in response to regular changes inday-length and temperature (Cramp and Simmons 1977 ; delHoyo et al. 1992 ). Seasonal freezing and thawing of wetlandscan also induce large-scale movements in the NorthernHemisphere. Winters in Australia are generally benign, withtemperatures rarely dropping below freezing point except insouthern mountains where there are few wetlands.

Most endemic Australian waterbirds are nomadic or known to move irregularly and, in the past, it was thought

L. E. Chambers ( * )Bureau of Meteorology Research Centre,G.P.O. Box 1289 Melbourne, Vic 3001, Australiae-mail: [email protected].: +613-9669-4784Fax: +613-9669-4660

R. H. LoynArthur Rylah Institute for Environmental Research,Department of Sustainability & Environment,P.O. Box 137 Heidelberg, Vic 3084, Australia


2/13

that waterbirds used the temporary wetlands of Australia sarid interior opportunistically from their preferred habitatsin more mesic regions of southern and eastern Australia (Roshier et al. 2001a ). For the ducks of Australia, as well asvarious other Australian birds that use the arid zone, thetiming and length of the breeding season can be quiteerratic, although underlying seasonal patterns are evident inmore temperate areas (Briggs 1977 , 1992 ; Halse andJaensch 1989 ; Marchant and Higgins 1990 ; Geering et al.1998 ; Roshier et al. 2001a ; Kingsford and Norman 2002 ).During summer in southern Australia, waterbirds tend toconcentrate in coastal regions with more regular and con-stant rains, and hence more permanent wetlands (Frith1982 ). Mass movements to coastal regions are also knownto occur during prolonged dry periods (Gentilli and Bekle1983 ; Gosper et al. 1983 ; Loyn et al. 1994 ; Kingsford and Norman 2002 ). One such coastal refuge is Western Port, acoastal embayment in Victoria, south-eastern Australia.This paper provides a case study documenting effects of climatic fluctuations over a 29-year period on the abun-dance of selected waterbird species in Western Port. The purpose of this paper is to explore the influence of broad-scale and local climate on the abundance of selectedwaterbird species in Western Port, over a longer time-framethan has been attempted previously for this or other Australian wetlands.

Studies, such as this one, provide quantitative data on theeffects of changes in climate on waterbird numbers and provide useful information for assessing how waterbirdsmay respond to future climate change. Recent studies sug-gest that many eastern Australian wetlands may experiencereduced rainfall, increased evaporation, and decreasedsurface-water flows in the future (e.g. Chambers et al. 2005 ).

Materials and methods

Study area

Situated to the south east of Melbourne, the large coastal bay known as Western Port (Fig. 1) covers an area of around 67,000 ha. Western Port is one of the most important wader and waterfowl sites in Australia, in terms of numbers of birds that it supports, and is listed under theRamsar Convention (Loyn et al. 2001 ). Since October 1973, regular counts of waterbirds have been made in thisregion by the Bird Observers Club of Australia, making it

one of the first large sites in Australia with a long record of waterbird abundance (Loyn 1978 ; Dann et al. 1994 ; Loynet al. 1994 , 2001 ; Heislers 2003 ).

Bird species considered

Three species were chosen for this study: the Black Swan(Cygnus atratus ), the White-faced Heron ( Egretta novae-

hollandiae ) and the Grey Teal ( Anas gracilis ). All threespecies are commonly seen in Western Port, and previousstudies indicate that they respond in different ways toseason and climate (Loyn et al. 1994 ).

The Black Swan has been defined as nomadic and proneto erratic dispersal (Pringle 1985 ). However, Marchant andHiggins ( 1990 ) note that when suitable habitat is per-manently available the adults are sedentary, but the young,and adults from ephemeral habitats, can move far andoften. The species breeds commonly on wetlands close toWestern Port and is classed as a local breeding species(Loyn et al. 1994 ).

White-faced Herons are found in Western Port through-out the year and some breed locally, although there is ageneral decrease during the spring breeding season (Loynet al. 1994 ). White-faced Herons are classified as sedentaryor locally nomadic, and may move away from the coast during the breeding season (Pringle 1985 ; Marchant andHiggins 1990 ).

The Grey Teal is widely distributed across Australia andgenerally considered to be nomadic, responding rapidly tochanges in rainfall and hydrology (Frith 1982 ; Gentilli andBekle 1983 ; Marchant and Higgins 1990 ; Loyn et al. 1994 ;Kingsford and Norman 2002 ). It is an important gamespecies in states where duck hunting is permitted (in-cluding Victoria) (Norman and Powell 1981 ; Briggs et al.1985a ; Loyn 1991 ). Few Grey Teal breed in the WesternPort region, and most breeding is believed to occur inephemeral swamps away from the coast.

Waterbird counts

Teams of observers made simultaneous counts of all

waterbirds on selected days, usually at high tide, at strategic sites around Western Port, including French andPhillip Islands (Fig. 1). The 17 most important sites for waterbirds were counted relatively consistently across sur-vey dates, and this paper focuses on data from these 17sites. From October 1973 to June 1976, counts were madeevery month with the day selected so that high tide fellwithin 4 h of midday and the day was on a weekend (Loynet al. 1994 ). Since June 1976, counts were made five times per year, twice in summer (December and February) andonce in autumn (April or May), winter (June, July or earlyAugust) and spring (September or October). The springand autumn counts were discontinued from 1996. Over 500

voluntary observers have taken part in the study since itsinception (Heislers 2003 ). The fieldwork is described morefully in previous papers (Loyn 1978 ; Dann et al. 1994 ;Loyn et al. 1994 , 2001 ; Heislers 2003 ). This paper usesseasonal count data from December 1973 to February 2002(127 dates). Summer count values were taken as themaximum of the December and February counts, but almost identical results were obtained when the mean of theDecember and February counts was used.

293


3/13

Allowance for variable coverage

The analysis was restricted to 17 sites that were included inthe regular survey and counted on most dates. Nevertheless,some of these sites were missed occasionally, for examplewhen high winds prevented the use of small boats. Al-lowance for variable coverage was made using an initialAnalysis of Variance, as follows. First, data were examinedfor each species across the 17 sites, with a view to excludingany sites rarely used by that species. This proved necessaryfor Grey Teal (6 sites retained), but not for Black Swan andWhite-faced Heron, which occurred commonly at most sites. Secondly, a two-way Analysis of Variance was con-ducted for each species, with date ( n=127) and site ( n=17

for White-faced Heron or Black Swan, 6 for Grey Teal) asfactors. This approach deliberately paid no attention toseasonal or climatic variables, or interactions between var-iables. Data were transformed to meet the assumptionsof this analysis (fourth root for Grey Teal; square root for White-faced Heron; no transformation for Black Swan). After appropriate back-transformation, tableswere generated for the means of each species for eachof the 127 dates, corrected for variable coverage. Thesecorrected means were then used as the basis for further analysis.

Climatic variables considered

Three types of climatic variables are considered, operating at different spatial scales. The Southern Oscillation Index(SOI) operates at broad global scales, affecting weather through the Pacific region (Allan et al. 1996 ). Rainfall datawere obtained for various regions of Australia, along withstreamflow data for selected catchments. All variables areexpected to influence the extent and productivity of wetlands in various parts of Australia (Norman and Nicholls1991 ; Roshier et al. 2001a , b; Kingsford and Norman 2002 ).

Rainfall and the Southern Oscillation Index (SOI)

Area-averaged (all Australia) monthly rainfall was ob-tained from the Australian Bureau of Meteorology s National Climate Centre. This index is a measure of howwet Australia is as a whole, with data available from 1890to the present (Jones and Weymouth 1997 ).

In addition, gridded monthly rainfall data were extractedfrom the Australian Bureau of Meteorology, NationalClimate Centre s 0.25 degree spatial data set (Jones andWeymouth 1997 ). Based on these data regional rainfallindices were also considered:

Fig. 1 Location of Western Port and other sites mentioned inthe text

294


4/13

1. Victoria - rainfall averaged over the state of Victoria2. Murray-Darling Basin - rainfall averaged over the

Murray-Darling Basin3. East Gippsland - averaged over the region 147.00E to

148.25E and 37.75S to 38.50S.4. Western District (SW) - averaged over 141.00E to

144.00E, 36.50S to 38.50S5. Western District (NW) - averaged over 141.00E to

143.00E, 34.00S to 36.00S6. Western Port - averaged over 145.00E to 145.50E,

38.00S to 38.75S.

Rainfall in these regions is thought to influence theabundance of some species of waterbirds in Western Port (Loyn et al. 1994 ), based on known movements and breeding distributions of waterbirds in south-easternAustralia (Norman 1971 ; Marchant and Higgins 1990 ;Barrett et al. 2003 ).

The Southern Oscillation Index (SOI) was calculated bythe National Climate Centre from monthly fluctuations inair pressure difference between Tahiti and Darwin (Troup1967 ).

Streamflow data

Monthly total volume (megalitres) streamflow data wereobtained, from the relevant state authorities, for a number of gauging sites around Australia (Fig. 2). The location of the stream gauges reflects the distribution of predominantly permanent or semi-permanent rivers and streams, most located along coastal areas and southern regions of thecontinent. No data were collected for the northern NewSouth Wales coastal region, Hydro Tasmania sites (south

western Tasmania), the Western Plateau division (wherethere were very few streams and even fewer available data),or the southern Northern Territory (where few data wereavailable). The drainage divisions and river basins areshown in Fig. 3.

Seven regions were considered, and streamflow aver-aged over all gauges in the region, in relation to waterbirdcounts in Western Port:

1. Gulf of Carpentaria (Division IX): 35 gauges used2. Northeast Coast (Division I): 74 gauges used3. Murray-Darling Basin (Division IV): 345 gauges used4. South (Division II - basins 14 39, Division IV - basins

1 15): 337 gauges used. (These drainage basins are inthe southern parts of the divisions; see Fig. 3)

5. South of Divide (Division II - basins 21 39): 135gauges used

6. Inland (Divisions X and XI): 7 gauges used7. Tasmania (Division III): 29 gauges used

These regions cover distinct climatic provinces over most of eastern Australia. This broad-scale approach is justified by the wide continental distributions of manywaterbirds, including the three species under consideration(Marchant and Higgins 1990 ; Barrett et al. 2003 ). Manycatchments cover vast areas, especially in arid Australia,and may receive water from thousands of kilometresdistant. Recent work has highlighted that ephemeralfloodwaters in arid Australia are key drivers of continental populations of many waterbird species (Kingsford et al.1999a ).

Although some streamflow data were available for thewestern divisions (VI, VII and VIII), waterbird populationsin these regions are generally considered to be isolatedfrom those in the eastern parts of the continent, with the

WesternAustralia

NorthernTerritory Queensland

New SouthWales

Victoria

Tasmania

South Australia

Fig. 2 Locations of streamgauge sites for which datawere obtained. Plottedover the region 10Sto 45S, 110E to155E

295


5/13

central arid zone inhibiting regular movement (e.g. Gentilliand Bekle 1983 ). Subsequent analysis of all availablestreamflow data (not shown) confirmed that streamflow ineastern Australia was more highly correlated with waterbird numbers in Western Port than streamflow in the west of the continent.

Correlations between the species counts were compared, both with and without adjustments for season in allanalyses.

Initially, relationships between the climate parametersand waterbird counts were investigated using correlationanalysis. To allow for possible seasonal variations incounts, independent of climate, counts for each seasonwere considered separately. As climate is unlikely to havean instantaneous effect on the number of waterbirdscounted in Western Port, particularly in more distant re-gions of Australia, lags of up to 16 seasons were consideredin the correlation analysis. This allows for the climate-waterbird system to have a memory of up to 4 years.

However, little credence was placed on correlations in-volving long time lags and distant regions, especially for those involving data from northern Australia, as most of theknown movements of the three species considered areconcentrated within the temperate and arid parts of Australia and not the tropical north (Marchant and Higgins1990 ). There are several reasons for considering a range of time lags (Briggs 1977 ; Norman and Nicholls 1991 ). Water can take many months to flow from the tropics or theGreat Dividing Range, where it falls as rain, to inlandwetlands where it provides breeding habitat for waterbirds(Kingsford et al. 1999a ). Waterbirds may use these habitats

for several years until they dry out, forcing birds to seek refuge elsewhere including coastal habitats such as WesternPort.

The seasonal count data were also combined for eachspecies to form an annual series with four values per year (summer, autumn, winter and spring). This data set was

analysed using General Linear Models with an indicator for season and a climate covariate (for example, All-Australian Rainfall). General Linear Models were used tolook for relationships between climate variables, such asrainfall, and waterbird counts, while taking into account the seasonal cycle in the count data. This allowed themodelling of both seasonal and climate effects on thewaterbird counts and gives an estimate of how much of the variation in the waterbird counts can be explained byseasonal and climate effects. Again, lags of up to 16seasons were considered.

Results

Waterbird count data: seasonal and year-to-year variation

There was considerable interseasonal and interannualvariability in the Western Port waterbird counts. Numbersof all three species showed distinct seasonal patterns( p


6/13

compared with 18.2% for White-faced Heron and 41.7%for Black Swan. Numbers of White-faced Herons andBlack Swans showed a positive correlation with each other (r =0.271, p=0.007) when considered on the basis of 127individual counts with no regard to season (reflecting thesimilar seasonal cycles for these species, with peaks inautumn and troughs in spring). Numbers of White-facedHerons and Grey Teal showed a weaker positive correlationwhen considered on the basis of standardised counts(allowing for seasonal effects) ( r =0.229, p=0.023). Noother significant correlations were found between seasonal

or annual counts of the three species. On an annual basisthere was little relationship between the species counts.Because the seasonal effect was so strong for two of thethree species, subsequent analyses are based on generalisedlinear models after allowing for the effects of season.

Black Swans were particularly numerous in 1973 1977and 1994 1995. White-faced Herons were most numerousnear the start of the survey, particularly in 1976. Grey Tealshowed an erratic pattern of peaks, with the major onesoccurring in 1975, 1977, 1982, 1992 and, to a lesser extent,2001. The lowest annual averaged counts of Grey Tealoccurred in 1974, 1988, 1996 and 2000.

Relationships of waterbird counts with climatic data

All three species showed significant relationships with theSOI (Tables 1, 2, 3), regional rainfall (Tables 1, 2, 3) andstreamflow (Tables 4, 5, 6), with appropriate seasonal lags.In each case, streamflow explained more of the count variance than the SOI or regional rainfall. Streamflow dataexplained up to 10.2% of the variance for Black Swan,14.5% of the variance for White-faced Heron and 34.5% of the variance for Grey Teal, having allowed for seasonaleffects in each case (Tables 4, 5, 6). The SOI was the

poorest predictor of waterbird numbers, explaining amaximum of 2.8% of the count variance for Black Swan,9.1% of the variance for White-faced Heron and 4.7% for Grey Teal (Tables 1, 2, 3). Regional rainfall explained up to6.2% of the variance for Black Swan, 12.2% for White-faced Heron and 9.9% for Grey Teal, having allowed for seasonal effects in each case (Tables 1, 2, 3). The SOI,regional rainfall and streamflow predictors performed quitesimilarly to each other for White-faced Heron, but verydifferently for the other two species (with SOI and regionalrainfall being markedly inferior to streamflow data). Theserelationships are explored further below.

B l a c k S w a n

SpringWinter AutumnSummer

10000

a

c

b

8000

6000

4000

2000

0

W h i t e - f a c e d

H e r o n


1000

800

600

400

200

0

G r e y

T e a l


700

600

500

400

300

200

100

0

Fig. 4 Box plots of season counts for each of the three species, showing median values, the interquartile range (values within the box , themiddle 50% of all counts) and extremes. a Black Swan, b White-faced Heron, c Grey Teal

297


7/13

SOI - Waterbird relationships

The SOI for the current season made a negative contribu-tion to models for numbers of Black Swans ( p=0.032), but the relationship was weak and explained only 2.8% of variance (Table 1). SOI for 1 3 seasons before the count made positive contributions to models for numbers of White-faced Herons, explaining 9.1% of variance with the

3-season lag ( p=0.001; Table 2). SOI for 5

8 seasons before the count made positive contributions to models for numbers of Grey Teal, explaining 4.7% of variance withthe 7-season lag ( p=0.030; Table 3).

Similar results were obtained when counts for eachseason were considered separately (not shown), with thespring species counts having the strongest relationshipswith the SOI (for up to six seasons lag). When the SOI washigh, spring waterbird counts tended to be low in all threespecies, particularly the Grey Teal and White-faced Heron.

Rainfall - Waterbird relationships

Rainfall data from various parts of Victoria made signif-icant though weak contributions to models for numbers of

1970 1975 1980 1985 1990 1995 2000

0

2000

4000

6000

8000

10000

B l a c k

S w a n

1970 1975 1980 1985 1990 1995 2000

0

200

400

600

G r e y T e a

l

1970 1975 1980 1985 1990 1995 2000

0

200

400

600

800

1000

W h i t e

. f a c e

d H e r o n

1970 1975 1980 1985 1990 1995 2000

-2

0

2

4

B l a c k

S w a n

A n o m a l y

1970 1975 1980 1985 1990 1995 2000

-1

0

1

2

3

4

G r e y

T e a

l A n o m a l y

1970 1975 1980 1985 1990 1995 2000

-2

0

2

4

W h i t e - f a c e d

H e r o n

A n o m a l y

a

b

Fig. 5 a Adjusted species counts over time. b Seasonal anomalies (seasonal mean subtracted and result divided by seasonal standarddeviation) of species counts over time

Table 1 Partial correlations between Black Swan count data and seasonal values of the Southern Oscillation Index (SOI) or regionalrainfall; after accounting for season

Lag SOI Rainfall

All-Australia Victoria Murray-Darling East Gippsland Western Dist. (NW) Western Dist. (SW) Western Port

0 0.219 # 0.131 0.223 # 0.160 0.256 # 0.186 0.197 0.1761 0.158 0.157 0.053 0.060 0.038 0.064 0.078 0.0472 0.137 0.094 0.100 0.038 0.151 0.071 0.080 0.1963 0.157 0.097 0.018 0.106 0.068 0.047 0.026 0.1084 0.094 0.087 0.068 0.091 0.014 0.081 0.100 0.0205 0.016 0.033 0.202 # 0.107 0.226 # 0.168 0.136 0.252 #

6 0.043 0.005 0.264 # 0.061 0.254 # 0.193 0.229 # 0.324 ##

7 0.024 0.042 0.122 0.024 0.091 0.093 0.095 0.1738 0.034 0.045 0.135 0.047 0.029 0.151 0.128 0.1029 0.114 0.000 0.287 # 0.148 0.180 0.261 # 0.283 # 0.295 #

10 0.174 0.071 0.051 0.056 0.044 0.053 0.045 0.056

Lag is the number of seasons lagged, for example, 0 is simultaneous and 1 uses the previous season s rainfall. Hashes indicate partialcorrelations significant at the 5% level; a single # indicating a percentage of variance explained by the SOI or rainfall in the range 2.5 5.0%;## in the range 5.1 10.0%; and ### over 10%

298


8/13

Black Swans (Table 1). Weak negative relationships werefound for rainfall in the season of the count, and positiverelationships with rainfall five or six seasons before eachcount (Table 1). Rainfall data from Western Port sixseasons before each count made a positive contribution,explaining 6.2% of the variance ( p=0.001; Table 1).

For White-faced Herons, plausible models could bedeveloped using rainfall data from a broader range of regions including All-Australia (Table 2). Relationshipswere generally positive for rainfall seven to nine seasons before each count. Rainfall data from All Australia sevenseasons before each count made a positive contribution,explaining 12.2% of the variance ( p


9/13

Streamflow waterbird relationships

For the Black Swan, streamflow in the region South of Divide (the smallest region incorporating Western Port) upto six seasons prior to the count made positive contributionsto the models and explained 3% to around 10% of thevariance in the counts (Table 4). Streamflow in the broader South region (which included most of Victoria and parts of southern New South Wales), three to six seasons prior to thecount, made similar positive contributions (Table 4). The best relationship was with data from the South region sixseasons before each count, explaining 10.2% of variance( p


10/13

Discussion

Seasonal patterns

The study showed a marked difference in consistency of seasonal pattern between the three species: Black Swanscould be described as having a strong seasonal cycle,White-faced Herons a moderate seasonal cycle and GreyTeal a weak seasonal cycle. Gosper et al. ( 1983 ) also foundthat counts of Black Swans were more seasonal than thoseof White-faced Heron and Grey Teal, in a 4-year study of waterbirds in the Richmond Valley, New South Wales. Theerratic movements of Grey Teal are well known: this highlymobile species breeds mainly in shallow swamps created by erratic inland flooding (Frith 1982 ; Gentilli and Bekle

1983 ; Marchant and Higgins 1990 ). Grey Teal disperseover long distances once these swamps dry out, appearingin coastal regions of Australia and New Zealand (Pringle1985 ; Marchant and Higgins 1990 ). Grey Teal are rapiddispersers with individuals capable of covering 180 km per day and up to 3,200 km in a few weeks (Roshier et al.2001a ). They are usually the first ducks to reach a newlyflooded area (Pringle 1985 ). Minimum numbers of all threespecies tended to occur in spring in Western Port, or winter and spring for Black Swan. This period corresponds to themain breeding season for many waterbirds in south-easternAustralia (Frith 1982 ; Marchant and Higgins 1990 ). Thetwo locally breeding species have been observed breeding

in nearby farmland and swamps, outside the study area, inthese respective seasons (spring for White-faced Heron;winter and spring for Black Swan) (Loyn et al. 1994 ).Fewer birds are seen during counts when they havedispersed to breed in nearby wetlands. Similar conclusionshave been reached from other studies in south-easternAustralia, most of which have shown that minimumnumbers of waterbirds occur on specific wetlands duringspring (e.g. Gosper et al. 1983 ; Kingsford and Norman2002 ). Peak numbers of waterbirds have been reported onlarge waterbodies in south-eastern Australia, at various

times from late summer through autumn into winter (Frith1982 ; Gosper et al. 1983 ; Marchant and Higgins 1990 ).

Grey Teal can breed at almost any time of year whenwater is available (Frith 1982 ; Marchant and Higgins1990 ), but in southern Australia wetlands are most likely tofill and provide good breeding conditions in spring (Dexter et al. 1986 ; Briggs and Thornton 1999 ; Roshier et al.2001b ; Kingsford and Norman 2002 ; Wright 2004 ). This isa consequence of high rainfall in winter and spring in muchof south-eastern Australia, along with spring snow-melt from the high mountains. Hence the average abundance pattern for Grey Teal showed similar spring minima to theother species (Fig. 5), but it varied far more between years(Fig. 5). Loyn et al. ( 1994 ) considered that inland breedingspecies, such as the Grey Teal, are less likely to have a

stable seasonal pattern in abundance between years thanlocal breeders, but they also observed a similarity in av-erage patterns when data were combined across manyyears.

Relationships with climatic variables

Many of the climatic variables were inter-correlated, andobserved relationships with waterbird numbers do not necessarily provide strong evidence that birds wereresponding directly to a particular region, rather thananother region with similar rainfall patterns. Interpretation

needs to consider both the movements of the birds and themovement of the water, both of which are extensive at thecontinental scale (Frith 1982 ; Kingsford and Norman2002 ). For example, we observed correlations betweennumbers of Grey Teal and rainfall in the Gulf of Carpentaria in far north Queensland, over 3,000 km awayfrom Western Port. Grey Teal are known to travel as far asthis (and further), but shorter movements are more com-mon (Norman 1971 ). A more plausible mechanism for such correlation lies with the movement of water from thetropical north during the summer wet season into inland

Table 6 Partial correlations between Grey Teal count data and seasonal values of the regional streamflow; after accounting for season

Lag Gulf of Carpentaria NE Coast Inland Murray-Darling South South of Divide Tasmania

0 0.014 0.048 0.082 0.241 ## 0.211 # 0.276 ## 0.1181 0.033 0.056 0.116 0.137 0.151 0.180 0.0042 0.074 0.024 0.124 0.067 0.083 0.028 0.1093 0.016 0.004 0.054 0.334 ### 0.405 ### 0.153 0.1344 0.107 0.097 0.009 0.252 ## 0.308 ## 0.152 0.1455 0.324 ### 0.133 0.179 0.159 0.142 0.008 0.220 #

6 0.596 ### 0.380 ### 0.545 ### 0.093 0.040 0.195 0.0217 0.137 0.145 0.040 0.121 0.098 0.114 0.1148 0.019 0.013 0.014 0.012 0.003 0.079 0.0179 0.009 0.011 0.131 0.046 0.105 0.034 0.13910 0.060 0.144 0.040 0.018 0.047 0.015 0.048

Lag is the number of seasons lagged, for example, 0 is simultaneous and 1 uses the previous season s streamflow. Hashes indicate partialcorrelations significant at the 5% level; a single # indicating a percentage of variance explained by the streamflow in the range 2.5 5.0%; ## in the range 5.1 10.0%; and ### over 10%

301


11/13

floodplains where Grey Teal breed in large numbers for short periods (Kingsford et al. 1999a ). Rainfall data fromthe Gulf of Carpentaria may act as a surrogate for rainfallgenerally in the tropical north. Water actually used by GreyTeal is more likely to have fallen as rain further south ontropical parts of the Great Dividing Range, within thecatchments of vast inland floodplains and river systems.

Australia s weather systems are driven by global air andwater currents, and the Southern Oscillation Index (SOI)has proved a useful indicator of long-term patterns (Allanet al. 1996 ; Scott 1997 ). Some relationships were found inthis study between the SOI and waterbird numbers, as inother studies in south-eastern Australia (Norman and Nicholls 1991 ; Roshier et al. 2001a , b; Kingsford and Norman 2002 ). However, in this study it did not prove asgood a predictor of waterbird numbers in Western Port asregional rainfall or streamflow data. This may be becausethe SOI is an imperfect and spatially coarse indicator of weather patterns that actually eventuate. Indeed, the dataused in this study showed that the influence of the SOI onrainfall in southern Australia was generally less strong thanit was for rainfall in other regions (Fig. 6), reflecting theseasonal predictability of rainfall in temperate southernAustralia in comparison with more northern or arid regions.The best predictors for all three species examined were thestreamflow data, which integrate effects of previous rainfallat the level of specific catchments, despite imperfectionsand anomalies arising from artificial regulation of many of the rivers. Wetland processes and waterbird numbers mayrelate more directly to streamflow data than rainfall data, because much rain feeds terrestrial and groundwater sys-tems with little impact on wetlands used by waterbirds.Streamflow data have also been used in other studies (e.g.Briggs and Holmes 1988 ). Loyn et al. ( 1994 ) used River

Murray flows as an integrating index of recent wetness inthe Murray-Darling Basin. Briggs and Holmes ( 1988 )developed a wetness index for various regions of Australia,and Roshier et al. ( 2001a , b; 2002 ) developed a process for estimating wetland areas from satellite imagery. Theseapproaches clearly have great potential for further under-standing waterbird dynamics at local and continental scales.

The present study confirms expectations from previouswork that waterbirds in Western Port are least commonwhen there is plentiful water in breeding habitats for therespective species (locally or inland) and most common at various periods of time after wet periods in those habitats.The study quantified those lag periods as three to nine

seasons (up to 4 years) for the three species considered inthis time-span (1973 2002).Conservation planning is typically conducted on a local

basis, recognising the values of local habitats and perhapstheir immediate context in the local landscape. But water-fowl appear to respond to climatic events at a broad con-tinental scale, and need habitats in widely different placesand times for different purposes. These fundamental pointsare well recognised by waterfowl biologists (e.g. Frith 1982 ;Marchant and Higgins 1990 ; Briggs 1992 ; Kingsford et al.1999b ; Roshier et al. 2001a ; Kingsford and Norman 2002 ) but perhaps less so by the general public or the planning

profession. The present study supplements the body of knowledge on these matters, and illustrates how numbers of birds in one area (Western Port) may be influenced byclimatic events in disparateand distant parts of the continent.These responses are only possible if suitable habitat hasbeen

conserved in these disparate areas. Ephemeral swamps are particularly important as breeding habitat for Australianwaterbirds for many reasons (Kingsford and Norman 2002 ).Such swamps provide rich supplies of invertebrate food for waterfowl, often based on decaying terrestrial vegetation asthe primary source of organic matter (Briggs et al. 1985b ;Crome and Carpenter 1988 ). Waterfowl are able to access alarge share of the available food, with few fish to competefor food or prey on ducklings. Many ephemeral swampssupport stands of trees such as River Red Gums Eucalyptuscamaldulensis , providing nest-sites on branches and inhollows (Briggs and Thornton 1999 ; Loyn et al. 2002 ).

Fig. 6 Simultaneous correlations between seasonal rainfall and theSouthern Oscillation Index (SOI) for a winter (June July August)and b spring (September October November), using data from1950 to 1999. The contour interval is 0.1 with shading starting at 0.3. Solid lines indicate positive correlations and dashed linesnegative correlations

302


12/13

Ephemeral swamps of this sort can support large numbers of breeding waterfowl, whether they are situated in temperate parts of Australia or far inland (Roshier et al. 2001a ).Conservation of ephemeral swamps must involve appropriate protection or management of vegetation and water regimes (Kingsford and Norman 2002 ). Maintaining water regimes is a special challenge in view of their enormousnatural variation in the arid zone, the complicating effects of climate change and the competing demands for water insettled parts of Australia. This challenge will intensify ashuman populations increase.

Acknowledgements The data on waterbird numbers could not have been collected without the dedicated efforts of numerousvoluntary observers over 29 years. Many thanks to them all, and tothe Bird Observers Club of Australia for supporting the survey.

Streamflow data were kindly provided by a number of stateauthorities: Department of Infrastructure, Planning and Environment (Northern Territory); Department of Primary Industries, Water andEnvironment (Tasmania); Department of Natural Resources andMines (Queensland); Department of Water, Land and BiodiversityConservation (South Australia); Rubicon Systems Australia Pty Ltd.(Victoria); Sydney Water (New South Wales); Water and RiversCommission (Western Australia). Mike Matheson, of the Bureau of Meteorology, provided assistance in obtaining information on streamflows, Wasyl Drosdowsky for the provision of the SOI-rainfall plots,Drte Jakob for assistance in producing the drainage division plot,and the National Climate Centre provided the SOI and rainfall data.

The authors wish to acknowledge the insightful comments made byDr Ian Norman of the Arthur Rylah Institute for EnvironmentalResearch, and Drs Scott Power and Bertrand Timbal, both of theBureau of Meteorology, on an earlier draft of this paper. The paper was further improved by comments from two anonymous reviewers.

References

Allan R, Lindesay J, Parker D (1996) El Nino Southern Oscillation

and Climate Variability. CSIRO Publishing, Collingwood,AustraliaBarrett G, Silcocks A, Barry S, Cunningham R, Poulter R (2003)

The New Atlas of Australian Birds. Royal AustralianOrnithologists Union, Hawthorn East

Briggs SV (1977) Variation in waterbird numbers at four swamps onthe northern tablelands of NSW. Aust Wildl Res 4:301 309

Briggs SV (1992) Movement patterns and breeding characteristicsof arid zone ducks. Corella 16:15 22

Briggs SV, Holmes JE (1988) Bag sizes of waterfowl in New SouthWales and their relation to antecedent rainfall. Aust Wildl Res15:459 468

Briggs SV, Thornton SA (1999) Management of water regimes inRiver Red Gum Eucalyptus camaldulensis wetlands for waterbird breeding. Aust Zool 31:187 197

Briggs SV, Maher MT, Davey CC (1985a) Hunter activity and

waterfowl harvests in New South Wales, 1977

1982. Aust Wildl Res 12:515 522Briggs SV, Maher MT, Tongway DS (1985b) Dry matter and

nutrient loss from decomposing Vallisneria spiralis L. Aquat Bot 22:387 392

Chambers LE, Hughes L, Weston MA (2005) Climate change andits impact on Australia s avifauna. Emu 105:1 20

Cramp S, Simmons KEL (eds) (1977) The Birds of the WesternPalearctic, Vol. 1. Oxford University Press, Oxford

Crome FHJ, Carpenter SM (1988) Plankton community cycling andrecovery after drought - dynamics in a basin on a flood plain.Hydrobiologia 164:193 211

Dann P, Loyn RH, Bingham P (1994) Ten years of waterbird countsin Western Port, Victoria, 1973 83. II. Waders, Gulls and Terns.Aust Bird Watch 15:351 365

Del Hoyo J, Elliott A, Sargatal J (eds) (1992) Handbook of the Birdsof the World, Vol. 1. Lynx Editions, Barcelona

Dexter BD, Rose J, Davies N (1986) River regulation and associatedforest management problems in the River Murray red gumforests. Aust For 49:16 27

Frith HJ (1982) Waterfowl in Australia. Angus & Robertson,Sydney

Geering DJ, Maddock M, Cam G, Ireland C, Halse SA, Pearson GB(1998) Movement patterns of Great, Intermediate and LittleEgrets from Australian breeding colonies. Corella 22:37 46

Gentilli J, Bekle H (1983) Modelling a climatically pulsating population: grey teal in south-western Australia. J Biogeogr 10:75 96

Gosper DG, Briggs SV, Carpenter SM (1983) Waterbird dynamics inthe Richmond Valley, New South Wales, 1974 77. Aust WildlRes 10:319 327

Halse SA, Jaensch RP (1989) Breeding seasons of waterbirds insouth-western Australia - the importance of rainfall. Emu89:232 249

Heislers A (2003) Wings over Western Port. Bird Observers Club of Australia, Nunawading

Jones DA, Weymouth G (1997) An Australian monthly rainfalldataset. Technical Report 70. Bureau of Meteorology, Mel- bourne, Australia, pp 19

Kingsford RT, Norman FI (2002) Australian waterbirds - products of the continent s ecology. Emu 102:47 69

Kingsford RT, Curtin AL, Porter J (1999a) Water flows on Cooper Creek in arid Australia determine boom and bust periods for waterbirds. Biol Conserv 88:231 248

Kingsford RT, Wong PS, Braithwaite LW, Maher MT (1999b)Waterbird abundance in eastern Australia, 1983 92. Wildl Res26:351 366

Loyn RH (1978) A survey of birds in Westernport Bay, Victoria,1973 74. Emu 78:11 19

Loyn RH (1991) Assessing and managing the impact of duck hunting in Victoria - a new approach. Wildfowl 42:155 161

Loyn RH, Dann P, Bingham P (1994) Ten years of waterbird countsin Western Port, Victoria, 1973 83. I Waterfowl and largewading birds. Aust Bird Watch 15:333 350

Loyn RH, Dann P, McCulloch E (2001) Important wader sites in theeast Asian-Australasian flyway: 1. Western Port, Victoria,Australia. The Stilt 38:39 53

Loyn RH, Lumsden LF, Ward KA (2002) Vertebrate fauna of Barmah Forest, a large forest of River Red Gum Eucalyptuscamaldulensis on the floodplain of the Murray River. Vic Nat 119:114 132

Marchant S, Higgins PJ (eds) (1990) Handbook of Australian, NewZealand and Antarctic Birds, Volume 1, Part B. OxfordUniversity Press, Melbourne

Nicholls N (2001) Atmospheric and climatic hazards: improvedmonitoring and prediction for disaster mitigation. Nat Hazards23:137 155

Norman FI (1971) Movement and mortality of Black Duck,Mountain Duck and Grey Teal banded in South Australia,1953 1963. Trans Royal Soc S Aust 95:1 7

Norman FI, Powell DGM (1981) Rates of recovery of bands, harvest patterns, and estimates for Black Duck, Chestnut Teal, GreyTeal and Mountain Duck shot during Victorian open seasons,1953 77. Aust Wildl Res 8:659 664

Norman FI, Nicholls N (1991) The Southern Oscillation andvariations in waterfowl abundance in southeastern Australia.Aust J Ecol 16:485 490

Pearce RP (2002) Meteorology at the Millennium. InternationalGeophysics Series, Volume 83, Academic Press, London

Pittock B (2003) Climate Change: an Australian Guide to theScience and Potential Impacts. Australian Greenhouse Office,Canberra

303


13/13

Pringle JD (1985) The Waterbirds of Australia. Angus andRobertson and the National Photographic Index of AustralianWildlife, Sydney

Roshier DA, Robertson AI, Kingsford RT, Green DG (2001a)Continental-scale interactions with temporal resources mayexplain the paradox of large populations of desert waterbirds inAustralia. Landsc Ecol 16:547 556

Roshier DA, Whetton PH, Allan RJ, Robertson AI (2001b)Distribution and persistence of temporary wetland habitats inarid Australia in relation to climate. Austral Ecol 26:371 384

Roshier DA, Robertson AI, Kingsford RT (2002) Responses of waterbirds to flooding in an arid region of Australia, andimplications for conservation. Biol Conserv 106:399 411

Scott A (1997) Relationships between waterbird ecology and river flows in the Murray-Darling Basin. CSIRO Land and Water Technical Report No. 5/97, pp 41

Stevermer AJ (2001) Recent Advances and Issues in Meteorology.Oryx Press, Westport, Conn.

Troup AJ (1967) Opposition of anomalies in upper troposphericwinds at Singapore and Canton. Aust Meteorol Mag 15:32 37

Wright WJ (2004) Drought, dust and deluge: a century of climaticextremes in Australia. Bureau of Meteorology, Melbourne

304

Chambers-L.E._2006 - La influencia de la variabilidad climática sobre los número de tres especies...

Documents

Transcript of Chambers-L.E._2006 - La influencia de la variabilidad climática sobre los número de tres especies...