BIRDS in LANDSCAPES

Project background

Natural levels of movement of individuals are critical for maintaining populations in fragmented systems. Movement is necessary to recolonise empty habitat, ‘rescue’ declining populations, and prevent negative genetic impacts. Thus, movement between fragments is likely to be critical for persistence in fragmented landscapes. Population viability is enhanced by natural levels of exchange of individuals and genes between sub-populations. Although rarely tested, it is assumed that movement is closely aligned with structural connectivity (eg corridors, stepping-stones).

So considerable emphasis has been placed on increasing structural connection to increase population persistence. However, there is scant evidence that movement is dependent on structural connectivity. A major impediment here has been the logistic constraints associated with quantifying movement using traditional ecological techniques (eg mark-recapture, radio-tracking). The advent of a variety of powerful genetic analyses means that movement can now be feasibly assessed.

Molecular information contained in the genotype provides a proxy for mark-recapture data but without the need for recapture. The accumulation of movement over many generations can also be resolved, to examine responses in relation to landscape change. Demographic processes and population trajectories can also be estimated. The integration of molecular data with relative abundance data from observational surveys and measures of landscape structure from GIS analyses (ie ‘landscape genetics’) generates a powerful toolkit for evaluating population-level responses to landscape change.

This project examines the relationship between movement of individuals, landscape structure (amount and arrangement of native vegetation) and population size, for selected bird species, building on the major landscape-scale innovative research of Radford et al. (2005) & Radford and Bennett (2007), who demonstrated that many species decrease in abundance with decrease of the size of the remaining remnants (‘decliners’) but some maintain or even increase their population density in small remnants in fragmented landscapes (‘resistant’). A fundamental difference between ‘resistant’ species and ‘decliners’ may be their ability to move through fragmented landscapes.

We employed a novel research approach that coupled existing data on the incidence of bird species in landscapes with molecular techniques that quantify the extent of within-landscape and between-landscape movement, and population trends, or trajectories. 

We assembled a unique dataset on current population size (survey data), extent of movement at several spatial and temporal scales (molecular data) and population trajectories (molecular data) for selected species in landscapes that vary in the amount and pattern of native vegetation.

We quantified the movement and other population processes of selected bird species within and among landscapes, enabling us to identify landscape features that promote natural population functions at patch, landscape and bioregional scales, and to compare population function and trajectories among species that have contrasting responses to landscape-level habitat loss.

Field work 

The fieldwork lasted 2.5 years from November 2007 to March 2010 and included mist-netting, trapping and banding birds, taking morphological measurements and collecting small blood and feather samples for genetic analyses and tests of bird health. Our study "landscapes" (801kb) are situated in the beautiful box-ironbark forests of central Victoria: nine of them have experienced different levels of habitat loss and fragmentation, and the other three contain continuous native forest.

Our landscapes and sites are:  

Landscape Boundaries

Site Locations

Target species included resident ‘decliners’ Brown Treecreeper, Superb Fairy-wren, Grey Shrike-thrush, Eastern Yellow Robin, Spotted Pardalote, Dusky Woodswallow, Buff-rumped Thornbill, mobile ‘decliners’ Fuscus Honeyeater, Yellow-tufted Honeyeater, Brown-headed Honeyeater, sedentary ‘resistant species’ Striated Pardalote and Weebill, and mobile ‘resistant species’ White-plumed Honeyeater.

Thank you to our amazing volunteers!

We are indebted to 68 volunteers who made their way from different corners of Australia, UK, USA, Japan, Taiwan, Russia, Switzerland, France, and New Zealand to share their time, expertise and devotion for birds and to help us during long hours of the field work, without whom this work would be much less efficient, if even possible. We are also grateful to landowners who allowed us access to their properties and sometimes helped us with the fieldwork. Congratulations to our field coordinators, Alan Lill and Naoko Takeuchi, on the great success of the field program.

Landscape genetics

Analyses of genotypes allow us to explore population structure and patterns of dispersal within and between landscapes. Also, using a small sample of DNA from blood or feather tissue, the sex of the individual birds can be identified. This is particularly useful for honeyeaters and other species that are difficult to sex in the hand. This information can be used to explore patterns of sex-biased dispersal. In addition, genetic data help to answer other important questions about kin associations and social structures (does level of relatedness within patch increase with level of fragmentation?), mating systems (does it change in fragmented landscapes?), parentage and extra-pair paternity in fragmented landscapes.

A massive amount of lab work by Sasha Pavlova, Katherine Harrisson and Naoko Takeuchi resulted in over 4100 birds of 10 species being sex-typed and genotyped for 6-16 highly resolving nuclear markers (microsatellites and epics, see the list of length-variable markers [Word, 28kb] that were used for genotyping). This number includes nearly 3300 individuals sampled during the fieldwork as well as historic samples from individuals collected Australia-wide by Australian National Wildlife Collection. We thank Rustam Turakulov for help with marker development.

Nev Amos, in collaboration with our partners Matt White and Graeme Newell developed models of landscape permeability for a range of bird species using multiple lines of evidence (expert opinion, banding recovery data, observational surveys).

Nev used several approaches that identify multiple dispersal paths through the landscape to predict critical areas for maintaining or promoting connectivity within and between landscapes. These models (Amos et al. 2012) were used to predict a priori the genetic distances that the landscape genetics will measure.

Nev’s modelling process involved incorporation of prior knowledge on species’ dispersal through different land uses (=landscape “resistance” values). The end result is a map of potential dispersal “current”, and pairwise resistance matrix for the study sites and/or landscapes for each species. Pairwise genetic distances are then used to test if landscape “resistance” (isolation-by-resistance, IBR) explains genetic distances beyond what could be explained by geographic distances (isolation-by-distance, IBD) via partial Mantel’s tests. The expectations were: (1) IBD and IBR in sedentary decliners; (2) IBR in mobile decliners; (3) Isolation effects weakest in mobile and tolerant species. 

Genetic estimates of dispersal were used to test these landscape permeability models and to identify the landscape features that are likely to enhance functional connectivity at inter-patch, landscape and bioregional scales (Amos et al 2014). Path analyses were performed for each of the species, and impacts of habitat configuration and landscape features on bird mobility were explored.

Molecular ecology data on social organisation, population dynamics, juvenile/sex-biased dispersal, species-specific patch connectivity, were analyzed in a landscape context (Harrisson et al. 2012, 2013a, 2013b, Pavlova et al. 2012, and additional papers on full publication list).

Our landscape genetic models were compared to Radford and Bennett species models (patterns tied to structural connectivity) which allowed us to relate structural landscape connectivity to mobility (processes underlying functional connectivity) and answer important questions such as: do decliners have lower functional connectivity than resisters under similar levels of structural connectivity? This knowledge trigger re-evaluation of Radford et al's (2005) 30% vegetation cover threshold (below which population decline is observed for many woodland species) with molecular estimates (indices of population growth/decline, estimates of within-patch relatedness and inbreeding, levels of within-patch genetic variation) and estimation of extinction debt (time lags between habitat loss and local extinction). Thus we could test how reliable are studies based on occupancy of patches.  

Mito-nuclear discordance as a result of selection on a female-linked trait

Understanding speciation is challenging because drift and selection can act at different times, or differentially on parts of nuclear and mitochondrial genomes. Using markers from both genomes and bioclimatic variables, Pavlova et al. (2013) examined the evolution of an Australian bird, the eastern yellow robin Eopsaltria australis. In southeastern Australia, two divergent mitochondrial lineages occur east and west of the Great Dividing Range, perpendicular to latitudinal structure of nuclear genes. Moreover, mitochondrial haplogroups correlate with climatic gradients present in the region owing to the Great Dividing Range: maximum temperature of warmest month explaines significan t amount of variance in mitochondrial variation above and beyond variance explained by geographic position and distance. Thus, it appears that environmental selection on a female-linked trait (mitochondrial DNA itself, mito-nuclear interactions or genes on female-linked W-chromosome) is driving mitochondrial divergence in the presence of nuclear gene flow. All alternative scenarios to explain this striking discordance in landscape genetic patterning (such as stochastic mtDNA lineage sorting, vicariance and female philopatry) were ruled out.

Hernan Morales followed up the causes of this intriguing pattern in his PhD work (Morales et al. 2015, 2017a, 2017b and 2018. Read more about his work and publications on Hernan's web page.  Building on Hernán's PhD, Stephanie Falk completed her PhD on the mitochondrial ecophysiology of this very intriguing system. See more on the Eastern Yellow Robin Project website.

Landscape bioacoustics 

Landscape genetics can be further integrated with other data informing about population connectivity, such as those on song-type similarity across the landscape. Sasha Pavlova and Nev Amos, in collaboration with Kate Buchanan (Deakin University), Maria Goretskaia and Irina Beme (Moscow State University) explored genetic and acoustic connectivity of the Grey Shrike-thrush by testing prior models of landscape permeability with genetic and acoustic data. They found that, although habitat fragmentation does not detectably influence Grey Shrike-thrush’s dispersal, it impedes bird’s song-type sharing, thus impacting important population processes (Pavlova et al 2012). Read more on the Grey Shrike-thrush story in this newsletter (pdf, 180kb).

Integration of landscape genetics and spatial modelling

Integrated data on landscape genetics and ecological permeability were incorporated into recommendations for mitigating the impacts of habitat loss and fragmentation on species survival.

Apart from these mainstream questions other important issues that have potential to impact population persistence and the nature of human-impacted populations were explored. Here are some examples:

Impact of habitat fragmentation on level of chronic stress in bird populations

Fragmentation could lead to a significant increase in the number of natural stressors affecting a local population of animals and result in repeated and frequent stress responses and, eventually, chronic stress. Members of populations experiencing chronic stress are likely to have reduced survival probability and reproductive output, and populations a greater chance of becoming extinct. Linda compared morphometrics and blood stress indicators (such as haemoglobin, hematocrit and H/L ratio, the ratio of two types of leucocytes) across landscapes with different level of fragmentation. These and other results are published in Amos et al. 2013.

Immune-response analyses

During her PhD, Shandiya investigated the effects of habitat fragmentation on the genetic diversity of immune system genes (MHC class II genes) and blood parasite infection in Brown Treecreeper, Striated Pardalote and Spotted Pardalote. Read Balasubramaniam et al. 2013, 2016 and 2017 to learn about Shandiya's results.

Conservation Genetics

Helmeted honeyeater Lichenostomus melanops cassidix is a highly endangered subspecies of a common eastern Australian species yellow-tufted honeyeater. Over 55 years it has attracted a lot of conservation resources, as well as critique for being insufficiently distinct to deserve conservation efforts. In particularly, Schodde and Mason 1999 hypothesized that cassidix represents just the end of a cline. In Pavlova et al. (2014) we analysed genetic data for individuals sampled across the species’ range, as well as morphological data for two subspecies, cassidix and gippslandicus, sampled across the hypothesized cline. We confirmed that cassidix is morphologically distinguishable from all other subspecies, is genetically distinct at the level of allele frequencies, has lowest level of genetic diversity and smallest effective population size among all subspecies and that it has been evolving independently from its closest relative, gippslandicus, for over 4,000 years. We propose that cassidix evolved through local adaptation and drift in small population and that maintenance of cassidix into the future should preserve some feature diversity of the species. We also suggest that introduction of gene flow from gippslandicus into the breeding program could facilitate cassidix’s adaptive potential and persistence (see Harrrisson et al. 2016). Have a look at our 5-minute AudioSlides presentation. We are following up this work in the Genetic Rescue Project.

Downloads

The map of our study sites (pdf 268kb)

Updated presentation for the 2009 stakeholders meeting (pdf 19.5mb)

Project newsletters:

Autumn 2008 (pdf 626kb)

Winter 2008 (pdf 430kb)

 Spring 2008 - Summer 2009 (pdf 636kb)

 Autumn - Winter 2009 (pdf 3.1mb)

Spring 2009 - Autumn 2010 (pdf 2.66mb) 

 Winter - Spring 2010

PUBLICATIONS

The outputs and outcomes of this project are explained in >25 publications.  

Their management implications are summarized in:

Radford JQ, Amos JN, Harrisson KA, Sunnucks P and Pavlova A. (2021) Functional connectivity and population persistence in woodland birds: Insights for management from a multi-species conservation genetics study. Emu – Austral Ornithology https://doi.org/10.1080/01584197.2021.1903331

see also

Perspectives in conservation

Practical genetic management

Genomes in conservation

Mitonuclear interactions

Habitat connectivity and fragmentation

Climate change biology

Speciation and biodiversity

Ecology and ecology/evolution tools

Other papers cited on this paper

Radford, J.Q., Bennett, A.F., 2007. The relative importance of landscape properties for woodland birds in agricultural environments. Journal of Applied Ecology 44, 737-747.

Radford, J.Q., Bennett, A.F., Cheers, G.J., 2005. Landscape-level thresholds of habitat cover for woodland-dependent birds. Biological Conservation 124, 317-337.