Methods of Environmental Evaluation for Avian Conservation

Tomohiro Ichinose

Abstract
1. Study objective
The objective of the present study is to propose procedures of environmental evaluation for avian conservation. The reasons why birds were selected as indicator organisms are as follows: 1) Birds are often accepted by citizens as symbols of natural landscapes, and 2) species composition of avian communities reflects landscape change clearly. Multivariate analysis methods were used for avian data analyses.
Species composition of avian communities can be influenced by various factors. In the present study environmental factors such as habitat area, habitat isolation and vegetation structure were mainly considered, because they can be manipulated through landscape planning and environmental management. Other factors such as interspecific relationship and density-dependent population dynamics were partly studied, though they might influence avian species composition.
2. Spatial scale of analysis
The environmental factors mentioned above are different in their spatial scale: Biogiographical conditions are possible environmental factors and should often be considered in national or even in global scale. Quite local conditions such as vegetational microstructures and boundary effects can also influence avian communities. It is not appropriate to analyze these conditions at a time. In the present study three classes of spatial scale of environmental conditions, namely “macro-scale”, “meso-scale” and “micro-scale”, were proposed to arrange data analysis properly.
The objectives of the “macro-scale” analyses are to classify avian habitats and elucidate their distribution within study areas. The environmental factors influencing habitat quality are identified in the following “meso-scale” analysis. Spatial heterogeneity and its influence on avian communities within the recognized habitat patches are considered in the final “micro-scale” analyses.
3. “Macro-scale” analysis
A procedure for data analyses using grid cell data system were proposed to carry out habitat classification and mapping. Multivariate classification methods for ecological data were recommended as tools of data processing. The steps of analysis were as follows:
1) On the basis of the census data covering the whole range of the study area, a set of grid cell data of avian species composition is prepared.
2) The grid cells are classified based on the avian compositional similarity. Multivariate analysis for classification is suitable for this stage.
3) The environmental factors causing spatial variation of species composition are detected. They can be considered to influence avian communities in the study area. Discriminant analysis using environmental parameters (for example, ratio of a particular land use or land cover) as independent variables is recommended to perform this stage.
4) On the basis of the classification typical habitat types and their spatial distribution can be estimated.
5) The area where more detail surveys are needed can be selected based on the discrepancy of actual status and estimated status obtained in the discriminant analysis.
To test the validity of the procedure, two sets of grid cell data were analyzed. One contains species compositional information of avian communities in a wintering and breeding season in Tokorozawa City, Saitama Prefecture, Japan. The other also contains avian compositional information in a breeding season in Sayama City, Saitama Prefecture. The land use was more diverse in Sayama City than in Tokorozawa City, because of the lack of rice fields and large rivers in the latter. The results of TWINSPAN and discriminant analysis indicated that not only the largest woodland among the study area but also fragmented woodlots would be important avian habitats in Tokorozawa City. In Sayama City a riparian corridor along the largest river and agricultural fields were also regarded as avian habitats maintaining distinctive species composition.
4. “Meso-scale” analysis
The “meso-scale” analysis was arranged in order that the important types of habitat indicated in the “macro-scale” analysis could further be investigated. The following procedure is proposed.
1) Among the typical habitat types indicated in the “macro-scale” analyses, one type is selected. Species composition of avian communities is then surveyed in the habitat patches of the selected type.
2) Relationships between environmental conditions at the habitat patches and avian parameters such as species richness and the number of individuals are examined by regression analysis.
3) The patches are classified based on the avian species compositional similarity. The use of multivariate classification methods such as TWINSPAN is recommended.
4) Discriminant analysis using habitat parameters such as area of habitat patches and isolation indices as independent variables is carried out to identify the environmental factors causing spatial variation of avian species composition.
5) The steps 1 to 4 could be repeated for other habitat types, if needed.
To test the validity of the procedure, avian communities in the 20 fragmented woodlots in a suburban area were studied in the Tokorozawa City. The results showed that woodlot area (i.e., area of habitat patches) was the most important predictor of the number of individuals and the species richness of birds in each woodlot. Based on the species composition of birds in wintering season, the fragmented woodlots were classified into four types by TWINSPAN, namely core woodlots, woodlots surrounding core ones, isolated and small woodlots, and isolated and middle-sized woodlots. By canonical discriminant analysis, it was indicated that connectivity between woodlots and vegetation coverage in the shrub layer were also important factors influencing avian communities.
5. “Micro-scale” analyses
Many factors influencing heterogeneity within habitat are considered in the “micro-scale” analysis. A procedure shown below was proposed based on the point-census technique and multivariate data analysis.
1) Birds are censused by the point census method at selected sites, among which local environmental conditions such as vegetation structure and vegetation composition are diverse regardless of the commonness of habitat type.
2) Relationships between avian parameters and environmental conditions are examined by regression analysis.
3) Sites are classified based on the avian compositional similarity.
4) The environmental factors causing spatial variation of avian species composition are detected by discriminant analysis.
Three case studies mentioned below were carried out considering vegetation composition, vegetation structure and effect of neighboring land use.
1) Relationship between avian species composition and vegetational species composition
Woodlots are usually heterogeneous for their vegetational species composition and structure. Such heterogeneity may cause spatial variation of avian communities. In the first case study of “micro-scale” analysis, attention was focused on the relationship between avian species composition and vegetational species composition. At 100 sites in 9 large woodlots located in Tokorozawa City or Sayama City, birds were censused in breeding season. The woodlot area ranged from 14.5 ha to 200 ha. Vegetation composition was also surveyed. The results showed that vegetation types had some correspondence with avian community types, which indicate the possibility of influence of vegetational heterogeneity on avian species composition.
(2) Relationship between avian species composition and vegetation structure
In the second case study the influence of vegetation structure on birds was mainly studied. Twenty-four sites were located on two isolated woodlots in Tokorozawa City, of which area was 14.5 ha and 7.8 ha, respectively. Birds were censused in a wintering season at each site. Vegetation structure was also surveyed. The results showed that vegetation cover density correlated with the avian species richness in wintering season. It was also indicated that avian species composition had some relationship with vegetation cover density in the shrub and herb layers and total basal area of standing-dead trees.
(3) Effects of contacting land uses on small urban woodlots and parks
Avian communities in isolated small woodlots and urban parks in residential area in Tokorozawa City were studied during a wintering season to identify the factors that influenced the distribution of wintering birds. The results showed that the ratio of edge contacting with urban land use to all edge explained the spatial variation of avian species composition, which implies the importance of edge effect in determining avian species composition. The location of the observed bird individuals indicated that several species such as Emberiza spodocephala and Corvus macrorhynchos showed tendency to prefer particular layers of vegetation, which suggests that vegetation structure may also influence birds distribution.
6. Animal ecological factors
In the present study interspecific relationship was observed through predation on bird nests. Predation ratio on artificial bird nests was experimentally studied in three woodlots on the Sayama Hills. The result did not indicate high predation ratio on edges, though many preceding studies indicated that predation ratio was higher near edges than in inner areas. Furthermore, predation ratio was highest in the largest woodlot and lowest in the smallest woodlot, which was also the opposite pattern against the common knowledge. The result was interpreted based on the observation that the density of Corvus macrorhynchos, which is a major predator in the study area, was significantly higher in the largest woodlot than in the other woodlots. It was suggested that the distribution of predators most influenced predation pattern on bird nests.
7. Application of the proposed methods
The methods proposed in the present study can be applied to environmental management or landscape planning. “Macro-scale” analyses can help to establish regional conservation area. “Meso-scale” analyses can be used in landscape planning or land use planning. “Micro-scale” analyses can be applied to local habitat maintenance (for example, re-vegetation, landscape planting, or construction of buffer zones around habitat patches). Evaluation of ecological factors may play a role to complement the evaluation of environmental factors.

Keyword:
bird community, environmental evaluation, multivariate analysis, habitat isolation, vegetation structure, landscape planning