Advertisement

International Journal of Biometeorology

, Volume 62, Issue 5, pp 873–882 | Cite as

Calling phenology of a diverse amphibian assemblage in response to meteorological conditions

  • T. Lynette Plenderleith
  • Danial Stratford
  • Gregory W. Lollback
  • David G. Chapple
  • Richard D. Reina
  • Jean-Marc Hero
Original Paper

Abstract

The strong association between amphibian activity, breeding and recruitment with local environmental conditions raises concerns regarding how changes in climate may affect the persistence of species populations into the future. Additionally, in a highly diverse assemblage of anurans, competition for breeding sites affects the time and duration of activity, as species compete for limited resources such as water. Meteorological conditions are strong drivers of amphibian activity, so we assessed whether temperature, rainfall, atmospheric pressure and humidity were associated with the calling phenology of an assemblage of anurans in South East Queensland, Australia. We performed calling surveys and collected digital recordings at 45 ponds in an area known for high anuran diversity. We performed detection analyses to investigate the influence of 10 meteorological variables in detection of calling activity in 19 amphibian species. Our results suggest four breeding strategies in the assemblage: explosive summer breeders, prolonged breeders, opportunistic breeders and a winter breeder. Classifying these species into associations provides a framework for understanding how species respond to environmental conditions. Explosive breeders (i.e. species demonstrating short and highly synchronised breeding periods) were particularly responsive to temperature. Our findings help elucidate the breeding phenology of frogs and provide valuable information on their mating systems in native Australian forests. This study highlights the difficulties of surveying even common anurans. We highlight the importance of predictability and stability in climate and the vulnerability of species for which reproduction appears to require highly specific environmental cues.

Keywords

Anuran Detection analysis Frog communication Litoria Myobatrachidae Bufo 

Introduction

Amphibians are restricted in time and location of breeding due to their dependence on suitable environmental conditions and aquatic habitats. In ephemeral wetlands, resources can be patchy, and competition, high (Aspbury and Juliano 1998). Many anuran species avoid permanent ponds that have a higher density of predators by breeding in temporary pools after rain (Skelly et al. 1999, Welborn et al. 1996). For successful recruitment in ephemeral wetlands, amphibians must breed and their larvae develop and metamorphose before water bodies dry out. Therefore, amphibian reproduction must be carefully timed to maximise chances of offspring survival (Richter-Boix et al. 2011).

In addition to the endogenous hormonal cycles that control reproductive timing, extrinsic factors play an important role in breeding phenology (Duellman and Trueb 1986), in maximising the probability of breeding and recruitment success in temporally variable habitats. Photoperiod (Canavero and Arim 2009), lunar cycles (Grant et al. 2009), temperature (Harkey and Semlitsch 1988, Wells 2010) and periods of rainfall (Donnelly and Guyer 1994, Gottsberger and Gruber 2004) all affect reproductive activity in amphibians, and most species probably require more than one factor to prompt activity (Koch and Hero 2007, Lowe et al. 2015, 2016, Shuker et al. 2016).

Higher temperatures increase the ectothermic frog’s ability to call and act as a cue that conditions are suitable for reproductive activity (Ritke et al. 1992). Temperature is an important cue for anuran reproductive activity (Oseen and Wassersug 2002, Lowe et al. 2016) and developmental rates and survivorship of anuran eggs and larvae are dependent on temperature (Moore 1939, Wells 2010). However, warmer, shallower breeding grounds may evaporate quickly and strand tadpoles, so there is an ecological trade-off between warm, shallow and fish-free temporary pools and permanent ponds with resident predatory fish.

Explosive breeding in amphibians is thought to be driven by the advantages of non-permanent (therefore fishless) ponds, as a breeding ground with decreased predation risk (Kats et al. 1988, Hero et al. 1998). A disadvantage to ephemeral breeding sites is that for most species, development through metamorphosis must occur before the wetland dries (e.g. Newman 1992). This mechanism creates conditions which favour aggregate breeding during times of rainfall (Segev et al. 2010) and short developmental periods in some species of frogs (Newman 1992). However, other species of anurans call over several months and the cues for these prolonged breeders are different and potentially more complex or subtle. Prolonged breeding is associated with seasonal rainfall patterns (Steelman and Dorcas 2010). Understanding the environmental requirements associated with breeding provides an indication of the potential threats which may be associated with changes in these conditions.

We investigated the relationship between local weather conditions and the calling behaviour of male anurans in a diverse frog assemblage in a subtropical coastal area of South East Queensland, Australia. Inter- and intraspecific competition is presumed high in this study system due to high species richness and abundance within ponds and little is known of the activity patterns of the species therein. Specifically, we aimed to document the phenology of male advertisement calling activity and analyse its relationship with temperature, humidity, atmospheric pressure, rainfall and season variables. We sought to classify each species into a representative breeding strategy (sensu Saenz et al. 2006), which can be used to help understand the fundamental ecology of reproduction in a diverse amphibian assemblage.

Methods

Study site

Surveys were performed at 45 permanent and ephemeral ponds in Karawatha Forest Park (KFP, 27° 38’ S, 153° 04′ E), Brisbane, Queensland, Australia (Fig. 1). Karawatha Forest Park is 909 ha of mostly sclerophyll eucalypt forest with a grassy understory, interspersed with Melaleuca forest patches and wetlands with an herbaceous understorey. The forest patch is surrounded by suburban settlement and connecting corridor forest. Karawatha Forest Park was logged in places during the 1950s (Kordas et al. 1993) and multiple areas have been burnt. Arsonist-lit fires are still problematic throughout the park. The urban settlement matrix has encroached on Karawatha Forest and a wildlife bridge was built over the motorway in the northwest of the park to connect it to adjacent forest patches (Jones et al. 2011). The terrain of KFP is reasonably flat, with an elevation of approximately 40 m above sea level (approximately 32–100 m in elevation range throughout the park). Brisbane and KFP have a subtropical climate with warm, humid summers and mild winters (Fig. 2).
Fig. 1

Location of Karawatha Forest Park (KFP) within Brisbane City region, Eastern Australia. Calling frogs were recorded at KFP and weather data was obtained from the Bureau of Meteorology station at Archerfield Airport

Fig. 2

Mean weather conditions from the Bureau of Meteorology Archerfield Airport weather station. Minimum (dashed line) and maximum (solid line) temperature data 1939–2015 and mean daily rainfall 1929–2015, adapted from Bureau of Meteorology (2015)

Calling surveys

We surveyed each pond for calling anurans. At each pond, we listened for 2 min, recording each species heard calling. Daytime and night-time surveys were performed approximately weekly from July 2007 to March 2010 and October 2013 to April 2014. We scored frog species calling (1) or not detected (0). Surveys were restricted to 2 min, as the majority of calling species are detected within the first 1–2 min of each hour sampled (this study, Shirose et al. 1997).

Audio recordings

As the ponds at Karawatha are spatially clumped, we ran a Mantel correlogram to test for spatial correlation of community composition and chose the five ponds most independent of one another. To record anuran vocalisations during the final year of the study, we placed Olympus DM-520 Digital Voice Recorders with double headed microphones (custom built standard microphone inserts soldered to cables and microphone jacks and coated in heatshrink waterproofing, Jaycar) at five ponds in KFP. We buried the recorders in waterproof housing at the very edge of the pond basin and suspended the microphones from vegetation approximately 1 m from the ground immediately above the recorder. We pointed the microphone heads towards the centre of the pond at approximately 90° to one another. The ponds at which the recorders were placed varied from permanent to fleetingly ephemeral (filled for only a few weeks). Recorders were set to record three times a day (mid-afternoon, sunset and midnight) for 2 min from September 2013 to April 2014. Due to technical difficulties, such as microphone flooding or detachment, not all days were represented with recordings. The recordings were later listened to by an observer. On the one instance that a definitive identification of a call could not be made, it was not added to the dataset.

In addition to the pooled data, a comparison of results from both recording and audio surveys was performed. We separated the human-led surveys from the recorded data to assess the difference between the techniques.

Weather data

Daily precipitation (to 0.1 mm), minimum dewpoint (°C), maximum dewpoint (°C), minimum temperature (°C), maximum temperature (°C) and mean sea level pressure (hPa) were provided by the Bureau of Meteorology Archerfield Airport station (BOM Site number: 040211; 27.57° 34’ S, 153.01° 0’ N; elevation 19 m) approximately 8 km northwest of Karawatha Forest. As frogs may call in response to rainfall from the previous day (LP pers. obs.), we incorporated the total accumulated rainfall from 2 to 3-day periods, as well as rainfall from the preceding 24 h. Mean sea level pressure was provided in eight readings daily (3 h apart), from which the highest and the lowest for each day were used in the analysis.

Classification of reproductive modes

For each species, we considered the breeding season as the months in which they were heard calling. The explosive summer breeders were defined as species with short periods of activity, requirement of warmer temperature for activity and no breeding recorded in two or more of the winter months, including June or July. The prolonged breeders were considered as such due to their year-round breeding activity (undetected in no more than 1 month). Species that had significant factors affecting calling behaviour, with the exception of Uperoleia laevigata (smooth toadlet), which was only heard once throughout the study, but that did not apparently follow a seasonal pattern were considered opportunistic breeders. This included Uperoleia fusca (dusky toadlet), which called at higher temperatures during November through March, and also in July (2008). Winter breeders were species that were heard in June and July (mid-winter), but not January (mid-summer).

Statistical analysis

The aim of the analyses was to investigate what environmental factors influenced frog species detection. We used a single season model explained by MacKenzie et al. (2002) to quantify detection rate using environmental variables as covariates. Only sites that had detections of the focal species over the sampling history were included in the analysis. To remove the effect of emigration, only sites that were active within a season were included within that season. Seasons were categorised as Austral spring/summer (September–February) and autumn/winter (March–August). This removed variable occupancy, thereby allowing an analysis of detection.

Correlations between environmental variables were quantified using a Spearman’s rank correlation, with correlated variables (ρ ≥ 0.7) not included in the detection analysis. Weather data, effort, day of the year (1–365/366) and a dummy variable for spring/summer season status (one for Austral spring/summer, zero for autumn/winter) made up the suite of environmental variables. Effort was the number of survey visits per day for the human surveys or the number of recordings per day for the audio recordings. The variable summer season may have been ineffectual for species that were never detected in either season because of the non-inclusion of sites based on seasonal results. Quadratic relationships between detection and day of the year and maximum temperature were included because there may be optimum conditions for when species call. For example, some species may be more detectable at the start and end of the year or vice versa.

A modified form of Akaike’s information criterion (AICc) was used for selection of the logistic regression detection models. All discussion in this section about AIC can be referenced to Burnham and Anderson (2002). AICc is used to avoid over-fitting and using the number of parameters and sample size to adjust the AIC value. We used the midpoint of the number of sites and number of sampling occasions as n. Models with an AICc difference (∆ i ) of 2 were considered as being the best descriptors of detection. Program Presence version 11.8 (Hines 2006) was used for the detection analyses.

Results

Species composition

We detected 19 frog species at KFP, representing three families of anurans (Table 1). There were 11 Myobatrachidae species, seven Hylidae and one non-native Bufonidae (Rhinella marina, cane toad).
Table 1

Detection of calling males by species in Karawatha Forest Park and the total number of observations (from 167 survey days) from July 2007 to March 2010 and October 2013 to April 2014

 

Jan

Feb

Mar

Apr

May

June

July

Aug

Sept

Oct

Nov

Dec

Total discrete detections

Myobatrachidae

Crinia parainsignifera (P) (eastern sign-bearing froglet)

X

X

X

X

X

X

X

X

X

X

25

Crinia signifera (P) (common eastern froglet)

X

X

X

X

X

 

X

X

X

X

16

Crinia tinnula (U) (wallum froglet)

X

 

X

 

X

 

X

X

X

X

25

Limnodynastes peronii (P) (striped marsh froglet)

X

X

X

X

X

X

X

X

X

X

35

Limnodynastes terraereginae (E) (northern banjo frog)

X

X

X

X

X

 

 

X

X

 

15

Platyplectrum ornatum (E) (ornate burrowing frog)

X

X

X

 

  

 

X

X

X

14

Pseudophryne major (W) (major toadlet)

 

X

X

X

X

X

X

 

X

X

11

Pseudophryne raveni (P) (copper-backed brood frog)

X

X

X

X

X

X

X

X

X

X

119

Uperoleia fusca (O) (dusky toadlet)

X

X

X

 

 

X

  

X

X

11

Uperoleia laevigata (U) (smooth toadlet)

 

X

  

  

    

1

Uperoleia rugosa (O) (wrinkled toadlet)

X

X

  

 

X

  

X

X

9

Hylidae

Litoria brevipalmata (E) (green-thighed frog)

X

X

  

  

 

X

X

X

15

Litoria caerulea (E) (green tree frog)

X

X

  

  

 

X

X

X

10

Litoria dentata (E) (bleating tree frog)

X

X

X

 

  

 

X

X

X

15

Litoria fallax (P) (eastern dwarf tree frog)

X

X

X

X

X

X

X

X

X

X

106

Litoria gracilenta (E) (dainty tree frog)

X

X

X

X

  

 

X

X

X

25

Litoria latopalmata (E) (broad-palmed frog)

X

X

  

  

 

X

X

X

7

Litoria rubella (E) (desert tree frog)

X

X

  

  

 

X

X

X

14

Bufonidae

Rhinella marina (P) (cane toad)

X

X

X

X

X

 

X

X

X

X

33

Months in which the frogs were detected are marked “X”; “–“indicates months not surveyed. Parenthetical codes refer to the reproductive group to which they were assigned: explosive breeders (E), prolonged breeders (P), winter breeder (W), opportunistic breeders (O) and unassigned (U)

The frequency of calling activity varied widely amongst species. The least heard species, U. laevigata, was only heard calling on one occasion (the night of the 6 February 2008), and at the other extreme, Pseudophryne raveni (copper-backed brood frog) was detected on 119 nights. Crinia parainsignifera (eastern sign-bearing froglet), Limnodynastes peronii (striped marsh frog), P. raveni and Litoria fallax (eastern dwarf tree frog) were all heard in every month of the year surveyed (Table 1). On the 29 October 2007, 15 species (all except Pseudophryne major (major toadlet)), U. fusca, U. laevigata and Uperoleia rugosa (wrinkled toadlet) were calling, but on 23 (of 167) survey nights, no frogs were heard calling. July was the quietest month, with only seven species heard calling, and November was the only month in which all species called. We classified the KFP assemblage of anurans into four modes of reproduction based on detection trend over time: explosive summer breeders, prolonged breeders, opportunistic breeders, and winter breeders (Table 1).

Pseudophryne raveni (detected on 119 surveys) and L. fallax (recorded on 106 surveys) were detected for the longest periods throughout the year, having been detected in every month in which surveys took place (Table 1). Conversely, U. laevigata was only detected once (6 February 2008), and Litoria latopalmata (broad-palmed frog) was only detected seven times, all throughout summer months (29 October 2007, three times in November 2008, once in December 2007 and once each in January and February 2008).

Detection analyses

The correlation matrix indicated a close relationship between all rain variables (Table 2), the two measures of dewpoint, dewpoint and daily minimum temperature and the two measures of pressure. Hence, the following variables were included in the detection analyses: daily rainfall, minimum temperature, maximum temperature, minimum pressure, day of the year, summer season and effort.
Table 2

Half matrix of Spearman’s rank correlations between environmental variables

 

DR

2DR

3DR

MinT

MaxT

MinD

MaxD

MinP

2DR

0.77 (0.00)

       

3DR

0.74 (0.00)

0.96 (0.00)

      

MinT

0.04 (0.80)

0.16 (0.27)

0.20 (0.17)

     

MaxT

− 0.38 (0.01)

− 0.15 (0.32)

− 0.13 (0.39)

0.47 (0.00)

    

MinD

0.27 (0.06)

0.41 (0.00)

0.45 (0.00)

0.75 (0.00)

0.38 (0.01)

   

MaxD

0.18 (0.21)

0.31 (0.32)

0.34 (0.02)

0.78 (0.00)

0.54 (0.00)

0.87 (0.00)

  

MinP

0.01 (0.97)

− 0.01 (0.97)

0.03 (0.85)

− 0.37 (0.01)

− 0.40 (0.01)

− 0.44 (0.00)

− 0.57 (0.00)

 

MaxP

− 0.01 (0.95)

− 0.07 (0.64)

− 0.05 (0.74)

− 0.42 (0.00)

− 0.40 (0.01)

− 0.54 (0.00)

− 0.59 (0.00)

0.91 (0.00)

ρ populates the cells with p value shown in brackets

DR daily rainfall, 2DR 2-day accumulated rainfall, 3DR 3-day accumulated rainfall, MinT minimum temperature, MaxT maximum temperature, MinD minimum dewpoint, MaxD maximum dewpoint, MinP minimum mean sea level pressure, MaxP maximum mean sea level pressure

Calling surveys

Species were detected at 4–41 sites over 49 sampling occasions. Uperoleia laevigata was not included in the analysis because of low detection. A suite of 11 a priori models were used to explain frog detection (Table 3). However, the model allowing for survey-specific detection had far too many parameters to be considered, thereby reducing the suite to 10. Constant detection over the survey period was the best performing model for Crinia tinnula (wallum froglet), L. latopalmata, P. raveni, R. marina, U. fusca and U. rugosa. Rainfall was the best predictor for four species, minimum temperature for three species, maximum temperature for two species, minimum pressure for two species and a quadratic form of maximum temperature for two species. Litoria caerulea (green tree frog) and Litoria dentata (bleating tree frog) were the only two species with competing models with an ∆ i  < 2. Species detection was positively associated with rainfall and minimum temperature, negatively associated with pressure and positively and negatively associated with maximum temperature, depending on the species (Table 4). Estimated average detectability from the best performing models varied from 0.73 for Pseudophryne raveni down to 0.26 for Litoria brevipalmata (green-thighed frog) and Litoria caerulea (Table 4).
Table 3

i of models for various species from data obtained from human-based surveys

Species

CDR

Rainfall

MinT

MaxT

MaxT2

MinP

TOY

TOY2

Season

Effort

Crinia parasignifera

15.42

14.49

16.86

0.00

2.51

16.75

17.50

18.45

16.76

14.70

Crinia signifera

5.29

7.19

5.30

0.00

2.43

4.04

7.73

3.49

5.51

7.31

Crinia tinnula

0.72

3.10

3.20

2.37

0.00

3.18

2.40

3.98

1.72

3.22

Limnodynastes peronii

27.09

0.00

26.90

29.15

29.01

29.31

29.34

29.96

29.38

26.67

Limnodynastes terraereginae

7.16

0.00

6.73

9.49

12.06

4.69

5.98

7.90

8.49

8.71

Litoria brevipalmata

16.70

7.92

5.26

15.45

9.00

0.00

14.02

15.61

19.12

19.02

Litoria caerulea

4.65

3.26

0.00

6.77

5.78

0.14

7.13

6.47

7.16

6.94

Litoria dentata

2.18

0.48

1.98

4.19

0.00

0.34

4.59

2.46

2.11

4.57

Litoria fallax

5.80

0.00

7.96

7.45

7.42

5.98

7.83

5.73

7.53

7.09

Litoria gracilenta

8.16

8.67

4.45

10.29

11.38

0.00

10.44

8.76

4.41

9.88

Litoria latopalmata

0.96

2.07

0.66

0.20

2.68

0.00

1.37

1.76

3.52

3.52

Litoria rubella

15.00

13.16

0.00

14.17

15.64

7.82

13.65

10.75

17.43

17.43

Platyplectrum ornatum

7.33

0.00

6.45

9.50

9.26

3.09

9.51

5.63

7.88

7.09

Psuedophryne major

21.38

19.98

5.69

7.05

0.00

12.41

23.32

1.17

7.78

22.44

Psuedophryne raveni

0.01

1.37

1.87

2.34

4.39

2.34

2.09

2.85

0.00

0.77

Rhinella marina

0.30

1.29

0.16

1.45

1.88

1.88

0.11

2.54

2.66

0.00

Uperoleia fusca

0.00

1.54

1.96

1.80

3.15

1.29

2.29

4.72

2.43

0.60

Uperoleia rugosa

0.00

1.39

1.89

2.36

4.45

2.36

2.11

2.91

0.02

0.79

Italic ∆ i indicates best performing model(s)

CDR constant detection rate, DR daily rainfall, MinT minimum temperature, MaxT maximum temperature, MaxT 2 maximum temperature in a quadratic form, MinP minimum mean sea level pressure, TOY time of year, TOY 2 time of year in a quadratic form, season summer season or not, effort number of surveys per day

Table 4

Best performing models for human-based surveys

Species

Best model formula

β 1 SE

β 2 SE

β 3 SE

Mean p

Mean SE p

n

Crinia parasignifera

logitp = 4.946 – 0.186 × MaxT

1.071

0.041

 

0.510

0.055

23

Crinia signifera

logitp = 3.672 – 0.164 × MaxT

1.520

0.058

 

0.355

0.074

16

Crinia tinnula

logitp = 0.000

0.378

  

0.500

0.095

9

Limnodynastes peronii

logitp = − 0.586 + 0.061 × Rainfall

0.207

0.015

 

0.520

0.044

41

Limnodynastes terraereginae

logitp = − 1.181 + 0.032 × Rainfall

0.288

0.012

 

0.329

0.057

19

Litoria brevipalmata

logitp = 219.448 – 0.218 × MinP

0.287

0.000

 

0.265

0.063

18

Litoria caerulea

logitp = − 8.034 + 0.359 × MinT

2.470

0.119

 

0.266

0.075

8

Litoria dentata

logitp = − 21.033 + 1.696 × MaxT − 0.034 × MaxT2

0.898

0.034

0.001

0.411

0.087

13

Litoria fallax

logitp = 0.500 + 0.033 × Rainfall

0.217

0.015

 

0.700

0.046

31

Litoria gracilenta

logitp = 112.000 – 0.111 × MinP

0.195

0.000

 

0.446

0.071

30

Litoria latopalmata

logitp = − 0.452

0.483

    

4

Litoria rubella

logitp = − 7.687 + 0.370 × MinT

1.671

0.083

 

0.364

0.061

17

Platyplectrum ornatum

logitp = − 1.247 + 0.051 × Rainfall

0.369

0.020

 

0.388

0.072

14

Pseudophryne major

logitp = 59.476 – 4.437 × MaxT + 0.080MaxT2

2.278

0.081

0.001

0.340

0.090

34

Pseudophryne raveni

logitp = 0.999

0.167

  

0.731

0.033

34

Rhinella marina

logitp = 0.378

0.184

  

0.594

0.044

28

Uperoleia fusca

logitp = − 0.660

0.362

  

0.364

0.084

11

Uperoleia rugosa

logitp = − 0.762

0.324

  

0.318

0.070

28

β parameter, p detection rate, n number of sites

Audio surveys

Acoustic recorders detected seven species of frog over 122 survey days. These species were C. signifera (eastern common froglet), C. tinnula, Limnodynastes peronii, Litoria fallax, P. major, P. raveni and U. fusca. Crinia signifera, U. fusca and U. major had identical detection histories and therefore had the same modelling results (Table 5). Time of year was the best performing model for all species except L. fallax, with effort being the best predictor of detection for this species. Detection rates varied from 0.71 for L. fallax to 0.17 for P. raveni (Table 6).
Table 5

i of models for various species from data obtained from human-based surveys

Species

CDR

DR

MinT

MaxT

MaxT2

MinP

TOY

TOY2

Effort

Crinia tinnula

19.53

20.95

21.72

19.62

21.15

9.38

8.32

0.00

14.97

Limnodynastes peronii

4.38

5.36

3.78

NA

NA

5.30

0.00

NA

5.33

Litoria fallax

21.60

18.16

23.64

22.15

17.55

17.96

23.54

19.90

0.00

Psuedophryne raveni

19.91

18.30

10.90

13.37

15.57

22.08

7.71

0.00

20.58

C. signifera, P. major, U. fusca

21.55

23.37

0.75

9.44

10.23

22.80

0.00

NA

23.75

Italic ∆ i indicates best performing model(s)

CDR constant detection rate, DR daily rainfall, MinT minimum temperature, MaxT maximum temperature, MaxT 2 maximum temperature in a quadratic form, MinP minimum mean sea level pressure, TOY time of year, TOY 2 time of year in a quadratic form, effort number of surveys per day, NA model did not converge

Table 6

Best performing models for human-based surveys

Species

Best model formula

β 1 SE

β 2 SE

β 3 SE

Mean p

Mean SE p

n

Litoria fallax

logitp = − 3.058 + 1.029 × Effort

0.71

0.246

 

0.447

0.035

5

Limnodynastes peronii

logitp = − 2.978 + 0.005 × Julian

0.33

0.001

 

0.179

0.036

3

Psuedophryne raveni

logitp = − 0.912 + 0.024 × Julian − 0.000 × Julian2

0.167

0.001

0

0.537

2.496

4

C. signifera, P. major, U. fusca

logitp = − 3.902 + 0.014 × Julian

0.505

0.002

 

0.468

0.065

1

Crinia tinnula

logitp = − 2.546 − 0.027 × Julian + 0.000 × Julian2

0.292

0.001

0

0.137

1.905

4

β parameter, p detection rate, n number of sites

Estimated detection rates can be used as a comparison between survey techniques. For C. tinnula, L. fallax, L. peronii and P. raveni, detection was much higher for the human-based surveys (0.50 vs. 0.14 for C. tinnula, 0.70 vs. 0.44 for L. fallax and 0.52 vs. 0.18 for L. peronii and 0.73 vs. 0.54 for P. raveni). Conversely, acoustic recorders detected C. signifera (0.47 vs. 0.36), P. major (0.47 vs. 0.34) and U. fusca (0.47 vs. 0.36) more often than human surveys. A comparison between the two methods is limited because of the difference in the number of sites and survey coverage over time.

Discussion

The explosive breeding group comprised L. terraereginae, P. ornatum, L. brevipalmata, L. caerulea, L. dentata, L. gracilenta and L. rubella. These species were positively associated with an increase in rainfall, minimum temperature or a decrease in atmospheric pressure. Most of these species had a clear best performing model, excluding L. dentata, who had four closely competing models, indicating a suite of environmental variables that trigger detection.

The prolonged breeders were L. fallax, L. peronii, C. parainsignifera, C. signifera, P. raveni and R. marina. The only species we considered to be a truly winter-breeding frog in the community was P. major, which was not detected in January or October and showed a negative relationship with maximum temperature, preferring not to call on hot days. All other species detected in both June and July were detected in every month of the year.

July is the coldest month on average (Bureau of Meteorology 2015) and was the month associated with least calling behaviour. January is the warmest month on average and only two species were not heard calling then—P. major (the winter breeder) and U. laevigata (which was only detected on one night during the whole study and was the only non-prolonged breeding species that was heard in all winter months surveyed). Crinia tinnula could not be categorised into any of the breeding categories and appeared to have a constant detection rate that did not vary with the measured weather conditions.

Weather was the best performing predictor of detection for 18 species, with all except L. dentata, having one clear best performing model. Of these species, rainfall performed best for four, temperature for five and pressure for two. It is important to note that dewpoint (a measure of humidity) was correlated with minimum temperature, the best predictor for two species. Rainfall provides much needed moisture for optimal physiological conditions such as metabolism, reduces desiccation (Fogarty and Vilella 2001, Hauselberger and Alford 2005, Woolbright 1985) and also fills breeding grounds. Pressure is often a precursor to rain.

Temperature, like moisture, can influence anuran metabolism and has previously been an important factor influencing anuran calling behaviour in another study from a subtropical area (Bertoluci and Rodrigues 2002). It appears to play a major constraining role in the two prolonged breeding Crinia species detected in this study and P. major, a winter breeder.

Nevertheless, some species like R. marina were robust to environmental conditions and displayed a constant detection rate. A constant detection rate was the best model for some species because of a lack of detection, meaning parsimony was most desired in model selection.

Temperate anuran species are often prompted to call by a combination of temperature and rainfall (Blair 1960, Canelas and Bertoluci 2007), whilst rainfall is usually the dominant driver of reproduction in tropical areas (Crump 1974, Donnelly and Guyer 1994, Toft and Duellman 1979) where duration is often controlled by other factors such as predators and resource availability (Wells 2010). Our results suggest that in subtropical Australia, temperature, rainfall and humidity are important. Many of the species have ranges that extend into temperate Australia and may therefore depend on temperature for reproduction due to phylogenetic constraints in the furthest extents of their ranges.

Summer rains are fairly reliable in the subtropics, and therefore, the warmer temperatures may signal a high likelihood of water-filled ponds, as well as being essential for metabolic processes and therefore activity. High humidity may be important as it indicates approaching rain and is integral to species that call out of the water, such as many Litoria species. Anecdotal evidence suggests that some species, such as L. dentata, only call during periods of intense and/or prolonged rain (LP pers. obs.). Whilst rainfall was an important predictor for this species, so were other environmental variables and the complexity of the relationship between calling of L. dentata and weather conditions remain unclear.

There was no evidence to suggest anuran succession in the assemblage at KFP. In each season, most species were detected on the same night as one another, as opposed to a turnover of species’ calling activity. This suggests that the anuran species of KFP use other anti-competitive strategies for successful co-existence or that resources are not limiting; neither of which was quantified.

Although reproductive patterns in frogs are on a continuum from one night to several months of breeding (Wells 1977), the frog species of KFP were mostly easy to divide into less than a month-long breeding period (explosive) to many months of breeding activity (prolonged). Additionally, the frog species of KFP represent explosive breeding strategies and prolonged strategies equally. Explosive breeding species were strongly affected by rainfall, and minimum and maximum temperatures were strong predictors of calling phenology in all breeding strategies employed at KFP. Both rainfall and temperatures in South East Queensland are expected to be altered by climate change (Cai et al. 2005). Air temperatures are projected to increase by 1.0 to 2.2 °C by 2050 and a decrease in rainfall is expected (Whitfield et al. 2010). Since most species at KFP rely on a combination of meteorological factors for their reproductive behaviour, the change in weather dynamics could make many of them vulnerable.

Some species may be harder to detect than others due to volume of their calls. In our study, nine species were not detected on the recordings, but were detected in human-led surveys. This could be due to the location of the recorders that supplemented the human surveys, but were only five in number. Regardless, the recorders likely did not have the same detection radius as the human surveys. We therefore recommend taking this into account when using audio recording equipment. Our comparison showed different results between recorded data and human-led survey data, which may be due to the lower detection rates of the audio recordings. The restricted time period of the acoustic recorders also played a role, highlighting the importance of surveying for more than one season. If a short time period within the breeding is chosen, it appears that sampling later in the summer season may increase detection, but this will depend on long-term weather conditions not measured in this study: a shortfall of the acoustic recorder study design within this study was the short length of time recorders were used.

Species that are only active in a limited range of conditions may be vulnerable to environmental change. Climate change could have serious negative effects on the breeding phenology of species that depend on a specific meteorological pattern (Beebee 2009, Blaustein et al. 2001, Klaus and Lougheed 2013). The significance of this is increased further for species such as L. brevipalmata, a species listed as endangered by the IUCN (Hero et al. 2004), which is also only active in a limited range of conditions. Further research is required to collect comprehensive data on the detailed requirements of each species to fully understand the implications of environmental change. We suggest recording weather conditions at a smaller scale than seen here and performing surveys most nights (more than 2 h post sunset) throughout the year.

Detection of amphibians is known to be problematic due to confounding environmental factors such as meteorological conditions. Furthermore, only four out of 18 species within this study had an estimated detection rate > 0.5. Hence, future studies should incorporate calculation of detection probabilities into study design to improve inferential rigour. To calculate detection probabilities, at least three repeat surveys are required.

The complex patterns of activity in response to meteorological conditions that we observed in KFP are likely reflected amongst amphibian assemblages globally. Amphibian populations are increasingly threatened (e.g. Stuart et al. 2004) and global climate change continues to increase its influence on weather patterns (IPCC 2014). Indeed, climate change is already implicated in amphibian declines (Pounds et al. 2006, McMenamin et al. 2008) and management plans must continue to consider environmental change whilst more research on the complex relationship between meteorological changes and amphibian activity is imperative.

Notes

Acknowledgements

We thank K. Buhagiar, C. Johnstone, M. Familiar-Lopez, C. Goulet and F. Hohaia for assistance in the field.

Funding information

The research was funded by the Ric Nattrass scholarship (Queensland Frog Society), Griffith School of Environment postgraduate support scheme, the Holsworth Wildlife Research Endowment Fund (grants to TLP) and a Science Faculty Early Career Researcher Grant (to DGC).

Compliance with ethical standards

This research was conducted under Monash University Biological Sciences Animal Ethics Committee approval BSCI-2013-20, Griffith University Animal Ethics Committee approval ENV/23/13/AEC and Queensland research and collection permit WISP14217514.

References

  1. Aspbury AS, Juliano SA (1998) Negative effects of habitat drying and prior exploitation on the detritus resource in an ephemeral aquatic habitat. Oecologia 115(1-2):137–148.  https://doi.org/10.1007/s004420050500 CrossRefGoogle Scholar
  2. Beebee TJ (2009) Amphibian breeding and climate. Nature 374:219–220CrossRefGoogle Scholar
  3. Bertoluci J, Rodrigues MT (2002) Seasonal patterns of breeding activity of Atlantic Rainforest anurans at Boracéia, Southeastern Brazil. Amphibia-Reptilia 23(2):161–168.  https://doi.org/10.1163/156853802760061804 CrossRefGoogle Scholar
  4. Blair WF (1960) A breeding population of the Mexican toad (Bufo valliceps) in relation to its environment. Ecology 41(1):165–174.  https://doi.org/10.2307/1931950 CrossRefGoogle Scholar
  5. Blaustein AR, Belden LK, Olson DH, Green DM, Root TL, Kiesecker JM (2001) Amphibian breeding and climate change. Conserv Biol 15(6):1804–1809.  https://doi.org/10.1046/j.1523-1739.2001.00307.x CrossRefGoogle Scholar
  6. Bureau of Meteorology (2015) Climate statistics for Australian locations. http://www.bom.gov.au/climate/averages/tables/cw_040211.shtml. Accessed 27 July 2015
  7. Burnham KP, Anderson DR (2002) Model selection and multimodal inference: a practical informationtheoretic approach, 2nd edn. Springer, New YorkGoogle Scholar
  8. Cai W, Crimp S, Jones R, McInnes K, Durack P, Cechet B, Bethols J, Wilkinson S (2005) Climate change in Queensland under enhanced greenhouse conditions, Report 2004-2005. CSIRO Marine and Atmospheric Research, VictoriaGoogle Scholar
  9. Canavero A, Arim M (2009) Clues supporting photoperiod as the main determinant of seasonal variation in amphibian activity. J Nat Hist 43(47-48):2975–2984.  https://doi.org/10.1080/00222930903377539 CrossRefGoogle Scholar
  10. Canelas MA, Bertoluci J (2007) Anurans of the Serra do Caraça, southeastern Brazil: species composition and phenological patterns of calling activity. Iheringia 97:21–26CrossRefGoogle Scholar
  11. Crump ML (1974) Reproductive strategies in a tropical anuran community. PhD thesis. University of Kansas, Lawrence, Kansas, USAGoogle Scholar
  12. Donnelly MA, Guyer C (1994) Patterns of reproduction and habitat use in an assemblage of Neotropical hylid frogs. Oecologia 98(3-4):291–302.  https://doi.org/10.1007/BF00324217 CrossRefGoogle Scholar
  13. Duellman WE, Trueb L (1986) Biology of amphibians. JHU Press, BaltimoreGoogle Scholar
  14. Fogarty JH, Vilella FJ (2001) Evaluating methodologies to survey Eleutherodactylus frogs in montane forests of Puerto Rico. Wildl Soc Bull:948–955Google Scholar
  15. Gottsberger B, Gruber E (2004) Temporal partitioning of reproductive activity in a neotropical anuran community. J Trop Ecol 20(03):271–280.  https://doi.org/10.1017/S0266467403001172 CrossRefGoogle Scholar
  16. Grant RA, Chadwick EA, Halliday T (2009) The lunar cycle: a cue for amphibian reproductive phenology? Anim Behav 78(2):349–357.  https://doi.org/10.1016/j.anbehav.2009.05.007 CrossRefGoogle Scholar
  17. Harkey GA, Semlitsch RD (1988) Effects of temperature on growth, development, and color polymorphism in the ornate chorus frog Pseudacris ornata. Copeia 1988(4):1001–1007.  https://doi.org/10.2307/1445724 CrossRefGoogle Scholar
  18. Hauselberger KF, Alford RA (2005) Effects of season and weather on calling in the Australian microhylid frogs Austrochaperina robusta and Cophixalus ornatus. Herpetologica 61(4):349–363.  https://doi.org/10.1655/04-03.1 CrossRefGoogle Scholar
  19. Hero J-M, Gascon C, Magnusson WE (1998) Direct and indirect effects of predation on tadpole community structure in the Amazon rainforest. Aust J Ecol 23(5):474–482.  https://doi.org/10.1111/j.1442-9993.1998.tb00755.x CrossRefGoogle Scholar
  20. Hero J-M, Hines H, Meyer E, Lemckert F, Newell D, Clarke J (2004) Litoria brevipalmata. The IUCN red list of threatened species. Version 2015.2. www.iucnredlist.org
  21. Hines JE (2006) PRESENCE—software to estimate patch occupancy and related parameters. USGS-PWRC. http://www.mbr-pwrc.usgs.gov/software/presence.html
  22. IPCC (2014) Climate change 2014: synthesis report. Contribution of working groups I, II and III to the fifth assessment report of the intergovernmental panel on climate change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, SwitzerlandGoogle Scholar
  23. Jones N, Bakker M, Bichet O, Coutts R, Wearing T (2011) Restoring habitat connectivity over the road: vegetation on a fauna land-bridge in south-east Queensland. Ecol Manag Restor 12(1):76–79.  https://doi.org/10.1111/j.1442-8903.2011.00574.x CrossRefGoogle Scholar
  24. Kats LB, Petranka JW, Sih A (1988) Antipredator defenses and the persistence of amphibian larvae with fishes. Ecology 69(6):1865–1870.  https://doi.org/10.2307/1941163 CrossRefGoogle Scholar
  25. Klaus SP, Lougheed SC (2013) Changes in breeding phenology of eastern Ontario frogs over four decades. Ecol Evol 3(4):835–845.  https://doi.org/10.1002/ece3.501 CrossRefGoogle Scholar
  26. Koch AJ, Hero J-M (2007) The relationship between environmental conditions and activity of the giant barred frog (Mixophyes iteratus) on the Coomera River, south-east Queensland. Aust J Zool 55(2):89–95.  https://doi.org/10.1071/ZO06017 CrossRefGoogle Scholar
  27. Kordas G, Coutts RH, Catterall CP (1993) The vegetation of Karawatha Forest and its significance in the South East Queensland landscape. Griffith University, NathanGoogle Scholar
  28. Lowe KJ, Castley JG, Hero J-M (2015) Resilience to climate change: complex relationships among wetland hydroperiod, larval amphibians and aquatic predators in temporary wetlands. Mar Freshw Res 66(10):886–899.  https://doi.org/10.1071/MF14128 CrossRefGoogle Scholar
  29. Lowe KJ, Castley JG, Hero J-M (2016) Calling phenology and detectability of a threatened amphibian (Litoria olongburensis) in ephemeral wetlands varies along a latitudinal cline: implications for management. Austr Ecol 41(8):938–951.  https://doi.org/10.1111/aec.12386 CrossRefGoogle Scholar
  30. MacKenzie DI, Nichols JD, Lachman GB, Droege S, Andrew Royle J, Langtimm CA (2002) Estimating site occupancy rates when detection probabilities are less than one. Ecology 83(8):2248–2255.  https://doi.org/10.1890/0012-9658(2002)083[2248:ESORWD]2.0.CO;2 CrossRefGoogle Scholar
  31. McMenamin SK, Hadly EA, Wright CK (2008) Climate change and wetland desiccation cause amphibian decline in Yellowstone National Park. Proceedings of the National Academy of Sciences 105(44):16988–16993Google Scholar
  32. Moore JA (1939) Temperature tolerance and rates of development in the eggs of Amphibia. Ecology 20(4):459–478.  https://doi.org/10.2307/1930439 CrossRefGoogle Scholar
  33. Newman RA (1992) Adaptive plasticity in amphibian metamorphosis. Bioscience 42(9):671–678.  https://doi.org/10.2307/1312173 CrossRefGoogle Scholar
  34. Oseen KL, Wassersug RJ (2002) Environmental factors influencing calling in sympatric anurans. Oecologia 133(4):616–625.  https://doi.org/10.1007/s00442-002-1067-5 CrossRefGoogle Scholar
  35. Pounds JA, Bustamente MR, Coloma LA, Consuegra JA, Fogden MP, Foster PN, La Marca E, Masters KL, Merino-Viteri A, Puschendorf R, Ron SR (2006) Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439(7073):161–167.  https://doi.org/10.1038/nature04246 CrossRefGoogle Scholar
  36. Richter-Boix A, Tejedo M, Rezende EL (2011) Evolution and plasticity of anuran larval development in response to desiccation. A comparative analysis. Ecol Evol 1(1):15–25.  https://doi.org/10.1002/ece3.2 CrossRefGoogle Scholar
  37. Ritke ME, Babb JG, Ritke MK (1992) Temporal patterns of reproductive activity in the gray treefrog (Hyla chrysoscelis). J Herpetol 26(1):107–111.  https://doi.org/10.2307/1565039 CrossRefGoogle Scholar
  38. Saenz D, Fitzgerald LA, Baum KA, Connwe RN (2006) Abiotic correlates of anuran calling phenology: the importance of rain, temperature, and season. Herpetol Monogr 20(1):64–82.  https://doi.org/10.1655/0733-1347(2007)20[64:ACOACP]2.0.CO;2 CrossRefGoogle Scholar
  39. Segev O, Hill N, Templeton AR, Blaustein L (2010) Population size, structure and phenology of an endangered salamander at temporary and permanent breeding sites. J Nat Conserv 18(3):189–195.  https://doi.org/10.1016/j.jnc.2009.08.006 CrossRefGoogle Scholar
  40. Shirose LJ, Bishop J, Green CA, Green DM, MacDonald CJ, Brooks RJ, Helferty NJ (1997) Validation tests of an amphibian call count survey technique in Ontario, Canada. Herpetologica:312–320Google Scholar
  41. Shuker JD, Simpkins CA, Lollback GW, Castley JG, Hero J-M (2016) Determining environmental limits of threatened species: the example of the wallum sedgefrog Litoria olongburensis. Ecosphere 7(6).  https://doi.org/10.1002/ecs2.1384
  42. Skelly DK, Werner EE, Cortwright SA (1999) Long-term distributional dynamics of a Michigan amphibian assemblage. Ecology 80:2326–2337CrossRefGoogle Scholar
  43. Steelman CK, Dorcas ME (2010) Anuran calling survey optimization: developing and testing predictive models of anuran calling activity. J Herpetol 44(1):61–68.  https://doi.org/10.1670/08-329.1 CrossRefGoogle Scholar
  44. Stuart SN, Chanson JS, Cox NA, Young BE, Rodrigues AS, Fischman DL, Waller RW (2004) Status and trends of amphibian declines and extinctions worldwide. Science 306(5702):1783–1786.  https://doi.org/10.1126/science.1103538 CrossRefGoogle Scholar
  45. Toft CA, Duellman WE (1979) Anurans of the lower Rio Llullapichis, Amazonian Peru: a preliminary analysis of community structure. Herpetologica:71–77Google Scholar
  46. Welborn GA, Skelly DK, Werner EE (1996) Mechanisms creating community structure across a freshwater habitat gradient. Annu Rev Ecol Syst 27(1):337–363.  https://doi.org/10.1146/annurev.ecolsys.27.1.337
  47. Wells KD (1977) The social behaviour of anuran amphibians. Anim Behav 25:666–693.  https://doi.org/10.1016/0003-3472(77)90118-X CrossRefGoogle Scholar
  48. Wells KD (2010) The ecology and behavior of amphibians. University of Chicago Press, ChicagoGoogle Scholar
  49. Whitfield L, Oude-Egberink K, Wecker B, Cravigan L, Pozza RD, Hernaman V, Scott J, Chidzambwa S (2010) Climate change in Queensland: what the science is telling us. Department of Environment and Resource Management,Brisbane 93Google Scholar
  50. Woolbright LL (1985) Patterns of nocturnal movement and calling by the tropical frog Eleutherodactylus coqui. Herpetologica:1–9Google Scholar

Copyright information

© ISB 2017

Authors and Affiliations

  1. 1.School of Biological SciencesMonash UniversityClaytonAustralia
  2. 2.School of Science and EngineeringUniversity of the Sunshine CoastMaroochydoreAustralia
  3. 3.CSIRO Land and WaterBlack MountainAustralia
  4. 4.Design Unit, EngineeringTweed Shire CouncilMurwillumbahAustralia

Personalised recommendations