Worldwide, 13 million hectares of forest are lost each year, mainly by conversion to agriculture.
This has raised concern about the loss of forest biodiversity in general and of habitat for forest
birds in particular; particularly in the biodiverse tropical forests of South and Central America.
Typically, some trees are retained within the agricultural matrix after conversion, and there is
mounting interest in the extent to which such landscapes, and the tree cover within them, can
provide habitat for birds. Here, this issue is explored by investigating the bird community using
the predominantly pastoral agricultural landscape of Rio Frio, in lowland Costa Rica. The
research involved mist-net sampling of birds in forest fragments, riparian forest, pastures with
trees, areas of young secondary forest, live fences and palmito (Bactris gasipaes) plantations
across the landscape. Data from 1850 captures of 120 species of birds, relocation of colourringed
individuals and spatial analysis of the tree cover from a satellite image were analysed to
address five key questions.
1. How does the bird community in the agricultural landscape compare with that of an adjacent
forest reserve? The bird fauna of Rio Frio was compared with that of an intact forest preserve at
La Selva Biological Station, which has the same life zone, forest type, altitude and climate. Bird
diversity was similar in the two places, with species richness at Rio Frio reaching 92 % of that at
La Selva, although there were considerable differences in bird community structure between the
two sites. Both resident and latitudinal migrant bird communities differed in terms of species
presence-absence and the dominance of shared species. Within Rio Frio, distinct bird
communities were associated with the different vegetation types sampled. The process of forest
conversion in Rio Frio appears to have led to the loss of many species in forest and understorydependent
families including Formicariidae, Troglodytidae, and Furnariidae, as well as largebodied
canopy dwelling species: but species richness and abundance of open-landscape species
have increased, including Tyrannidae, agricultural specialists and granivores.
2. Do the established forest-dependency classifications used for Costa Rican birds and based on
observations in forested landscapes, still apply in an agricultural landscape: or do species adapt
to the changed environment? The stability of established forest-dependency classifications for
Costa Rican bird species was investigated by comparing how birds used different vegetation
types in Rio Frio. Five established classifications from Stiles (1984) were reduced to three
functional groups based on birds' distinct patterns of vegetation use in this landscape. 'Forestobligate'
species used predominantly forest fragments and to a lesser extent palmito plantation.
'Forest-dependent' species dominated forest fragments but used some other vegetation types,
and 'forest-independent' species predominantly used agricultural matrix vegetation types. The
use of vegetation by most species accorded with the established classifications, but there were
interesting examples of some forest-dependent species extensively utilising agricultural
vegetation types; especially four understory species in palmito plantations. The study
highlighted how subdivision of the bird community into functional groups can be used to
identify which groups are most at risk in fragmented landscapes, and which vegetation types
within the agricultural matrix might be important in providing habitat for them.
3. Given the complexity of agricultural landscapes and their use by forest bird species as
already highlighted in 1 and 2 above, it was pertinent to explore whether the extent and
configuration of tree cover remaining after conversion affects the persistence of forest-dependent
bird species? This was done by comparing the bird communities using two contrasting parts of
the Rio Frio landscape. Forest-dependent birds formed a greater part of the bird community in
forest fragments than in the agricultural matrix, but were more common in a structurally
complex matrix than in a structurally simpler one. Larger forest fragments demonstrated an 'interior' bird community distinct from those in both small fragments and the edges of large
fragments. Forest fragments of more than 15 ha are irreplaceable resources for the most
sensitive group of forest-obligate and understory-dependent species that do not utilise dispersed
trees in the landscape matrix.
4 . What distances do birds typically travel in agricultural landscapes: how does this affect their
use of fragmented vegetation and hence the scales of description and analysis required to
understand interactions between birds and landscape structure? Scale of landscape use for
species in the Rio Frio bird community was measured by analysing inter-location distances for
703 encounters with 100 individual birds of 25 species, observed during 650 observer-days and
3500 km walked transects. Species means were grouped to give three significantly different
scales of landscape use: the smallest scale group (< 10 m) included all the forest-dependent
species studied except some Woodcreepers. Larger scale groups (55 m and 150 m) include
forest-independent birds and Woodcreepers that have increased in abundance or remained
stable following deforestation. These larger scales correspond to several vegetation patch sizes in
the Rio Frio landscape.
5. Does tree cover at the scale that bird species use the landscape affect body condition? This
was explored for three bird species in Rio Frio. For two forest-dependent species body condition
varied significantly and linearly with canopy cover in the agricultural matrix. Analyses at scales
inappropriate for the species reduced or completely obscured such relationships. Body condition
in the forest-independent species also varied with canopy cover but this relationship explained
only 50 % of the variation. Most, but not all of the measured variation in body condition for
forest-dependent species can be explained using image-based spatial analysis of canopy cover in
the agricultural matrix around the forest fragments in which they were captured. Increased
canopy cover in the agricultural matrix around forest fragments may have a positive influence
on the persistence of forest-dependent birds in fragmented landscapes like Rio Frio.
Rio Frio is a deforested agricultural landscape that supports high diversity in all bird groups,
including forest-dependent birds. Characteristics of this landscape here identified as valuable
for forest-dependent bird species were the remaining forest fragments, and a matrix structure
including retained trees which is structurally complex at a scale that makes several vegetation
types accessible to individual birds.
The study highlighted the value of all forest fragments to forest-dependent birds, even those as
small as 1 ha: with the caveat that the most vulnerable forest-obligate species were those least
able to utilise agricultural vegetation. Larger forest fragments contained a more specialised
forest bird community. The importance of tree canopy cover in the agricultural landscape was
identified by the effect of matrix complexity on forest-dependent bird species richness and
abundance: and underlined by its positive influence on body condition in birds, including that
of an understory species seldom encountered outside forest fragments.
These results suggest that the ability of the most vulnerable forest-dependent bird community
to use contemporary deforested landscapes can be increased by appropriate agricultural
practices. Retaining or allowing the regeneration of forest fragments, including fragments as
small as 1 ha and larger than 15 ha, could be combined with measures to increase canopy cover
and structural complexity in the agricultural matrix. Silvopastoral systems, replacing forest
trees in pastures when these are felled, the replacement of existing post-and-wire with live
fences and increasing the pollarding interval of such fences, all have the potential to encourage
the persistence of forest bird species outside forest preserves.