Corresponding author: Leonardo Fernandes França ( franca_lf@ufersa.edu.br ) Academic editor: Claudia Hermes
© 2020 Leonardo Fernandes França, Victória Helen Figueiredo-Paixão, Thales Afonso Duarte-Silva, Kamila Barbosa dos Santos.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
França LF, Figueiredo-Paixão V, Duarte-Silva TA, Santos K (2020) The effects of rainfall and arthropod abundance on breeding season of insectivorous birds, in a semi-arid neotropical environment. Zoologia 37: 1-7. https://doi.org/10.3897/zoologia.37.e37716
|
Rainfall in tropical semi-arid areas may act as a reliable cue for timing bird reproduction, since it precedes future food supply. With this in mind, we set-up a study to test the reproductive response of insectivorous bird to arthropod abundance and rainfall patterns. Sampling occurred in a seasonally dry Neotropical forest, in north-eastern Brazil, between October, 2015 and October 2016, at 14-day intervals. We used brood patch to assess reproductive periodicity of insectivorous birds (eight species, 475 captures, 121 patch records). We sampled arthropods to quantify abundance, using biomass and number of individuals (1755 individuals, 15 Orders). Rainfall temporal distribution was analyzed using daily precipitation data. We used a cross-correlation function to test for correlation and time-lags between the covariates under study. Both the number of reproductively-active birds and arthropod abundance were higher in time periods close to the rainy season. Increase in arthropod biomass in the aerial stratum preceded the period of greatest rainfall by one (14 days, r = 0.44) to three sampling periods (0.47). In contrast, the highest proportion of individuals with brood patches occurred after the main rainfall peak, with the strongest relationship occurring after two (0.52) to four (0.50) time lags. Finally, the proportion of individuals with brood patches was positively correlated with aerial stratum arthropod biomass when five time lags were considered (0.55). Our results support the hypothesis of a temporal process involving rainfall, arthropods and reproduction of insectivorous birds in the wet/dry tropics. However, rainfall did not appear to act as a cue for the timing of reproduction, since records indicated higher arthropod biomass before the main rainfall peak. At least occasionally in the study area, insectivorous bird reproduction peaks after food abundance.
Brood patch, Caatinga, food-mediated process, insect, reproduction
In tropical seasonally dry areas there is widespread evidence of links between rainfall, bird breeding phenology and clutch size (
The existence of specific food-mediated processes involving weather, food and bird reproduction were originally proposed for temperate regions, based largely on observed relationships between temperature and avian reproductive periodicity (
We explore the relation among rainfall-food-reproduction in an insectivorous bird assemblage in a seasonally dry tropical forest in north-eastern South America (
We conducted the study in Caatinga, a Neotropical dry forest type present in north-eastern Brazil (
We established a grid of four parallel transects, separated from each other by 100 m. Every 50 m along each transect we placed 10 sampling points, each point with a mist net (18 × 3 m, mesh 16 or 19 mm; 40 sampling points). Each track alternately received nets with a specific mesh width. Between October 2015 and October 2016, we sampled the area by opening mist nets at 05:00 a.m. and keeping them open until 10:00 pm. Sampling occurred every 14 days (27 sampling events), with the aim of capturing the speed of response of the organisms under study. Other studies have reported instances where the reproductive response of tropical birds and the increase in prey abundance occurred 10–14 days after environmental stimuli (
Captured birds were marked with coded metal rings provided by the Centro Nacional de Pesquisa para a Conservação das Aves Silvestres (CEMAVE/ICMBio – National Centre for Wild Bird Conservation). These birds were identified and categorized as reproductive by the presence/absence of a brood patch. When present, the brood patch was categorized as: (a) active patch – abdominal region lacking feathers, highly vascularized and with subcutaneous fluid present; or (b) inactive patch – abdominal region lacking feathers, but little vascularized, lacking subcutaneous fluid and with a dry and scaly appearance. For the data analysis we included only records of ‘active brood patches'. To quantify reproductive season we examined the active brood patch data and estimate the ratio between the number of individuals with brood patches and the total number of individuals captured during each sampling event. We combined males and females when performing the calculations as the species under study showed no apparent sexual dimorphism, except Formicivora melanogaster Pelzeln, 1868, which males also have brood patches. Methods for categorizing reproductive phenology followed those used in previous tropical bird studies (
Arthropods were sampled at three sites within the grid, all separated by 150 m and on the same days as bird sampling. Three sampling methods were used in each site: windowpane traps, pitfall traps and suction sampling (
To assess the abundance of food resources (arthropods) we used the number and biomass obtained by each sampling method and estimate proportions, using the ratio between abundance at each sampling event and total abundance recorded during the study. We combined data from windowpane traps and suction sampling and treated this as “aerial stratum arthropods". This was done because evidence indicates that, in Neotropical environments, aerial arthropod assemblages tends to suffer marked within-year variation in abundance, whereas abundance of leaf litter arthropods tends to be more constant (
Temporal distribution of rainfall was based on daily precipitation data collected with a rainfall gauge located at 10 km from the study site (data available in “CEMADEN – Centro Nacional de Monitoramento e Alertas de Desastres Naturais" – National Centre for Natural Disaster Monitoring and Alerting, http://www.cemaden.gov.br/mapainterativo). We calculated the cumulative precipitation for the 14 days prior to each sampling event, to test the short-term effect of this on the biological processes studied.
We restricted the analysis to eight species of birds whose diets consist of at least 90% invertebrates (diet descriptions:
We used a Cross-correlation Function (CCF) to test for correlations and time lags between accumulated precipitation, arthropod abundance and brood patch records. CCF was used to correlate two time series (y and x) to enable the effects of different time lags of the y-variable to be checked against the x-variable. Negative lags indicate that y-series values are related to previous x-series values, while positive lags indicate that y-series values precede y-series values (
During the study, rains occurred mostly between January and June, with annual precipitation (313 mm) below the regional mean. In arthropod traps we recorded 1755 individuals and identified 15 orders. From these, 1293 individuals were from the seven Orders considered in the analyses (Table
Aerial stratum arthropod biomass was positively correlated with rainfall distribution, preceding it by one to three time units (time unit = 14 days; Table
The two largest aerial arthropod biomass peaks occurred in final third of December and January, when in 28 days we captured 43% of the total aerial arthropod biomass recorded during the study. The highest accumulative precipitation occurred shortly after, between the end of January and February, when 55% of the total rainfall volume recorded fell in 39 days (Fig.
Abundance of arthropods from 27 sampling events, based on three sampling procedures (windowpane, pitfall, suction). Values correspond to the number of individuals and biomass (dry weight in grams). Data recorded between October 2015 and October 2016, with 14-day intervals, in an area of Neotropical seasonally dry forest in northeastern Brazil.
Taxon | Windowpane | Pitfall | Suction |
Insecta | |||
Coleoptera | 208 (0.786) | 215 (0.592) | 30 (0.108) |
Hemiptera | 2 (0.033) | 0 (0.00) | 0 (0.00) |
Hymenoptera | 119 (0.325) | 431 (2.439) | 18 (0.023) |
Isoptera (Infraordem) | 37 (0.619) | 1 (0.002) | 0 (0.00) |
Mantodea | 1 (0.001) | 0 (0.00) | 0 (0.00) |
Orthoptera | 5 (0.004) | 26 (0.151) | 50 (0.128) |
Arachnida | |||
Araneae | 12 (0.023) | 110 (0.357) | 21 (0.014) |
Larvae | 0 (0.00) | 1 (0.001) | 6 (0.013) |
Correlation coefficients (r) and time-lags of the cross-correlation function that had p < 0.05. Accumulated precipitation, arthropods abundance (number and biomass), and the proportion of individuals with active brood patches were correlated with each other. Each lag unit corresponded to a sample interval of 14 days, with up to six lags considered in each direction.
Covariables | r | lag | |
X | Y | ||
Accumulated precipitation | Biomass (aerial stratum) | 0.436 | 1 |
0.465 | 3 | ||
Accumulated precipitation | Biomass (pitfall) | – | – |
Accumulated precipitation | Number (aerial stratum) | – | – |
Accumulated precipitation | Number (pitfall) | – | – |
Accumulated precipitation | Biomass number (aerial stratum) | 0.394 | 3 |
Accumulated precipitation | Biomass number (pitfall) | – | – |
Biomass (aerial stratum) | Brood patch | 0.549 | -5 |
0.463 | -6 | ||
Biomass (pitfall) | Brood patch | – | – |
Number (aerial stratum) | Brood patch | – | – |
Number (pitfall) | Brood patch | – | – |
Biomass_number (aerial stratum) | Brood patch | 0.489 | -5 |
0.400 | -6 | ||
Biomass_number (pitfall) | Brood patch | – | – |
Accumulated precipitation | Brood patch | 0.462 | 0 |
0.430 | -1 | ||
0.521 | -2 | ||
0.382 | -3 | ||
0.501 | -4 |
(1) Biomass and (2) number of arthropods recorded during 27 sampling events. Solid lines represent the aerial arthropods (windowpane trap and suction sampler), while dashed lines show terrestrial arthropod (pitfall trap). Data recorded between October 2015 and October 2016, with 14-day intervals, in an area of seasonally dry Neotropical forest, in northeastern Brazil. The abundance (biomass and number) were converted to proportion using the ratio between abundance at each sampling event and total abundance during the study. *In one month, we had three, instead of two, monthly samplings as a result of the 14-day interval between samples.
(3) Aerial arthropod biomass, (4) accumulated precipitation and (5) active brood patches, during 27 sampling events. Data recorded between October 2015 and October 2016, with 14-day intervals, in an area of seasonally dry Neotropical forest in northeastern Brazil. The biomass were converted to proportion using the ratio between abundance at each sampling event and total abundance during the study. Brood patch was converted to proportion using the ratio between individuals with brood patches and total individuals captured during a sampling event. *In one month, we had three, instead of two, monthly samplings as a result of the 14-day interval between samples.
Our results support the hypothesis that there is a causative process involving precipitation patterns, food resources abundance (arthropod) and the reproductive period of insectivorous birds in the seasonally dry tropics (
However, we found no evidence for the food-mediated process described in previous studies, where the rainfall is used by birds as a cue to future food availability and so regulates the timing of breeding in a way that deals effectively with the temporal unpredictability of this resource (
It is possible that the studied birds used the short period of food surplus to gain enough energy reserves and then reproduce, interrupting the long period of food restriction imposed by the dry period. This approach runs against the prediction that food surplus in seasonal/unpredictable environments is used to feed nestlings and fledglings, reducing the impact of severe periods on their survival (
We suggest here the existence of alternative mechanisms to explain the reproductive timing of birds in dry tropical forests, such that the birds use food abundance rather than rainfall as a proximate cue to adjust the timing of reproduction. Although, explanatory power of the data was still limited by the temporal nature of this research and by the lack of information on other aspects of bird reproduction. For example, relationships between reproductive phenology and nest predation have yet to be clarified, as have the potential effects of factors that covary with food abundance or rainfall period (
We thank the graduate and post-graduate students of the Laboratory of Population Ecology at the Universidade Federal Rural do Semi-Árido (UFERSA) for their help with field work. This work was supported by UFERSA, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES, and Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq (undergraduate and postgraduate grants) and, CNPq (442524/2014-5) and Fundação de Apoio à Pesquisa do Rio Grande do Norte – FAPERN (005/2011/PPPIV/57) (research grants). Adrian Barnett revised the English version.