Determining the diversity of birds in Bornean tree plantations

DETERMINING THE DIVERSITY OF BIRDS IN BORNEAN TREE PLANTATIONS Alison Styring 1, Frederick H. Sheldon 2, Roslina Ragai 3 and Joanes Unggang 4 1 The Evergreen State College, Olympia, WA 98505, USA 2 Museum of Natural Science, Louisiana State University, Baton Rouge, Louisiana 70803, USA 3, 4 Grand Perfect Sdn Bhd, Bintulu, Sarawak, Malaysia ABSTRACT observers can then be trained quickly to identify indicators species, thereby providing accurate, cost To design and execute cost effective assessments of effective assessments of bird diversity. bird diversity in tree plantations is relatively easy given basic knowledge about bird communities and skills in identifying common species. Bird communities INTRODUCTION increase in diversity with the age and structural complexity of groves, and they are also likely to be The main reasons for surveying birds in industrial and influenced by other factors, such as proximity of agricultural plantations in Borneo are to improve our groves to natural forest, age of the plantation in terms understanding of the ecology of birds in tropical of crop rotations, and regional variation in indigenous forests and, then, to use this information to develop faunae. Surveys of birds should be designed to take management strategies to increase bird diversity in advantage of these factors and to assure consistency regenerating and artificial forests. The quest to among replicate plots. For example, a great deal may understand tropical forest bird ecology derives from be learned about bird diversity and community one of the great questions in ornithology: How do so development by comparing groves of different ages many species of birds coexist in tropical rainforests? (and thus structural complexity) as long as these Plantations offer an exceptional opportunity to groves also share inherent properties (e.g., adjacent examine this question because they comprise a natural habitats, soils, streams, cliffs, roads, logs, and snags). experiment in community development. The presence Accurate assessment of bird species diversity in a in a single plantation of different aged groves of trees, given location may be accomplished by the method of plus natural forest in buffers and surrounding areas, “distance sampling”. This method emphasizes the allows researchers to examine communities of birds in estimation of species density (number/area) and is different successional stages at a single point in time at accomplished by transect counts. Two kinds of data a single locality. By relating increases in bird diversity are collected during such counts, species with changing features of “aging” plantation forests, identifications of individual birds and estimates of ornithologists can discern habitat and community distance of individual birds from observers. From requirements of individual species. Wildlife managers these data, a wide variety of parameters may be can then translate this knowledge into strategies for estimated and inferences made using models designed plantation design and maintenance that encourage bird for distance sampling. With a little practice, observers diversity. can be trained to recognize many bird species and their songs and, thus, accomplish effective data From preliminary studies of bird communities in two collection. Moreover, in certain circumstances industrial tree plantations—the Grand Perfect “indicator species” can be identified through plantation in Bintulu, Sarawak, and the Sabah distance-sampling analysis. These species indicate Softwoods plantation near Tawau, Sabah (Mitra & pre-established levels of diversity. Thus, apprentice Sheldon, 1993)—we have already learned a great deal 134 Determining the Diversity of Birds in Bornean Tree Plantations about factors that influence bird community development and diversity. Bird diversity is correlated with forest complexity. Older plantation groves, with their high canopies and substantial understories of native plants, contain more bird species than younger groves, which are largely monocultures of plantation trees. This aging effect is most pronounced for plantation tree species and management strategies that encourage the development of complex secondary forests. For example, Albizia (Paraserienthis falcataria), with its unusually high canopy and light composite leaves, allows the greatest development of understory growth. Consequently, groves of Albizia tend to have the most diverse avian faunae. Acacia mangium is close behind, also because it permits substantial secondary forest development. On the other hand, oil palm (Alaeis guineensis), has the lowest diversity of birds of any plantation species we have examined. This is because oil palm fronds capture most of the in-coming light and allow virtually no understorey development. Other factors that influence bird diversity include proximity of plantation groves to natural forest; the closer to natural forest, the greater the influx of immigrant and commuter bird species. Also, regional differences can be important in determining diversity. The native forest in the region of Sabah Softwoods grows on rich volcanic soils and, as a result, its bird diversity is inherently greater than that found at Grand Perfect plantation, which has nutrient-poor sandy and peaty soils. Another factor we suspect will influence diversity is the overall age of plantations. Bird diversity is known to diminish through attrition and extinction as isolated islands of forest get older (Diamond et al., 1987). It is likely that as plantations age—and stumps and logs disappear, groves are cropped, and surrounding natural forest recedes—bird diversity will dwindle. Ficedula are rare or non-existent in logged forest and plantations (Wong, 1986; Lambert, 1992; Mitra & Sheldon, 1993), but we can only guess why this is so. In some cases, reasons for the relative abundance or rarity of bird groups are fairly obvious. A dearth of large canopy frugivores, such as pigeons, hornbills, and large barbets, in industrial tree plantations is readily explained by a lack of canopy fruits. Only rarely, however, do we possess an adequate understanding of the ecology of a particular group of birds to be able to predict its specific habitat requirements with accuracy. Such is the case with Malaysian woodpeckers. The niche parameters of individual species of woodpeckers have been examined through exacting research on ecology and morphology (Styring & Zakaria, 2004a & b), and as a result we know their habitat requirements very well and can predict when and where they will occur. Studies in tree plantations offer the opportunity to develop a similar level of understanding of other bird groups. Given the relationship between plantation structure and bird diversity, and the potential of plantations to provide critical information on bird community ecology and autecology, bird surveys should be designed not only to determine the number of birds that occur in plantations but also the environmental factors responsible for supporting those birds. Because of the natural experiment inherent in plantations, the design of information-rich surveys is fairly straight forward. Birds need to be counted in different aged groves of trees within the plantation and surrounding natural forest. If a plantation has plantings of different tree species, e.g., Albizia, Acacia mangium, Gmelina arborea, Eucalyptus deglupta, oil palm, etc., then bird occurrence in different tree plots should also be assessed. As birds are counted, so should their habitat be surveyed. It is extremely important to gather Although we know that bird diversity is roughly information on the forest composition and structure to correlated with forest complexity and age, we actually relate to bird diversity. understand very little about the specific habitat and life-history factors that influence individual species In this paper we provide details on the kind of data that and groups of birds. As a result, we generally cannot need to be collected to understand bird diversity in specify precisely which factors are responsible for Bornean plantations, and we describe how to collect increases in bird diversity, nor are we able explain why those data. We believe that much of this work could be some groups of birds decline in disturbed or artificial done by “paraornithologists” or “paraecologists”. We forests. For example, from plantation surveys, we have use the term paraornithologist as a parallel to the term learned that the three species of tailorbirds, Ashy “parataxonomist”, which refers to local people who, (Orthotomus ruficeps), Rufous-tailed (O. sericeus), by virtue of their knowledge of plants and animals, can and Dark-throated (O. atrigularis), tend to replace one contribute importantly to the assessment of another as groves age (Mitra & Sheldon 1993), but we biodiversity without formal academic training (Janzen have no idea which habitat characteristics determine 1993; Basset et al., 2000). Parataxonomists have been this trend. Also, we know that muscicapine flycatchers recruited in many countries to help document in such genera as Eumyias, Cyornis, Niltava and biodiversity by collecting and preserving museum 135 Alison Styring, Frederick H. Sheldon, Roslina Ragai and Joanes Unggang specimens. In a similar way, paraornithologists would undertake surveys of birds in Bornean plantations and forests and provide critical, low cost information on bird occurrence. We also believe, that plantations offer a tremendous opportunity for the training of undergraduate and graduate students at Malaysian universities. Such students could contribute substantially to our knowledge of bird diversity and conservation by pursuing research projects on the ecology of specific bird groups. The infrastructure for such research already exists in tree plantations (e.g., housing, roads, labor, silvicultural and botanical data, GIS technology, etc.). Thus, at relatively little cost, plantations provide an idea location to further our understanding of tropical bird ecology. METHODS “Distance Analysis” (Buckland et al., 2001) is currently considered to be the most comprehensive and accurate method for determining population characteristics of many groups of wildlife, including birds. This method depends mainly on the collection of two types of data: the number of birds detected and the distance of each bird from the surveyor. To determine numbers of birds occurring in an area, it is necessary to conduct a relatively large number of surveys, so that common species are counted accurately and all rare species are recorded. To assess density (i.e., individuals per area), which is the key parameter to estimating population size, it is important to estimate distances (hence, area covered) accurately. Once these data are collected, they can be analyzed using the program Distance (Thomas et al., 2003). The power of this program is that it estimates population size based on the detectability of species and, thus, controls for bias caused by habitat differences. For example, recently logged forest is more open and allows a greater range of visibility than primary or old secondary forest. Thus, birds can be detected from greater distances in logged forest than in other types of forest. If, during surveys, birds were simply counted in the different forest types, a larger number would be recorded in logged forest than other forests, whereas the actual number of birds would not necessarily be greater in logged forest. On the most basic level, Distance Analysis controls for habitat bias by weighting individuals observed at close distances more heavily than individuals at longer distances. In the program Distance, this bias is modeled with a detection function, which is simply the probability of detecting an individual at a given distance. The shape of this function will change depending on variables that influence detectability (such as forest type and species—some species are easier to detect than others), but it is generally assumed that as distance from the observer increases, detectability of individuals decreases. Therefore, individuals detected very far away from the observer add very little information or strength to the model. The models in Distance also take into account other factors, and as a result Distance can be used to determine a variety of survey parameters, such as the amount of sampling effort required to obtain accurate counts. To collect data for analysis in Distance, observers must conduct a series of surveys. Each survey consists of a transect of fixed distance during which birds are counted. Our plantation surveys, for example, consisted of 1 km transects divided into 20 points, each 50 m apart. At each point, we spent 3 minutes counting individual birds by sight and sound and measuring their distances from the point. Optimally, the count duration should be as short as possible to gain a relatively complete “snapshot” of the focal species in an area. The longer the duration of a count, the greater the chance of bias in population estimates due to bird movement. Because there are often multiple individuals and species vocalizing and moving through the habitat during a point-count, it may be difficult to focus on all the species present while estimating distances at the same time. Observers may want to construct a “map” of the survey point. This map is simply a bull’s-eye target drawing. The the middle represents the observer, and then several concentric outer circles indicate distances from the observer (Appendix 1). Before starting the survey, the observer locates easily recognizable landmarks in the count area (a large tree, snag, or stump, or the edge of a gap, etc.) and measures the distance from the observer to the landmarks. During the point-count, birds can be “mapped” (recorded) onto the bulls-eye according to their relative position. After the survey is completed, the observer can then measure distances using the “map” as a reference. Distances to the birds must be measured as radial (ground level) as opposed to line-of-sight distances. Thus, the distance to a bird 50 m high in a tree is measured from the observer to the tree trunk, not to the bird. Distances should be recorded with the aid of a measuring device. Styring and Ickes (2001) used 50-m tape when conducting surveys of woodpeckers at Pasoh Forest Reserve. This method of distance estimation was quite accurate, but time consuming. Tilt-compensated laser rangefinders, which may be found in any hunting or forestryequipment catalog, are the best choice for distance estimation because they are easy to use. You just aim the rangefinder at the bird and push a button. The tilt of the rangefinder adjusts the line-of-sight distance to 136 Determining the Diversity of Birds in Bornean Tree Plantations radial distance. The observer must also be able to measure distances between counting points. We strongly recommend the use a hand-held GPS unit for this purpose because it not only provides information on distance between points, but also allows the collection of georeferenced data that can be analyzed in GIS. Appendix 2 is a sample datasheet for bird surveys using distance sampling methods. To understand the relationship of bird abundance and community structure with habitat characteristics, observers must collect habitat data at each point where a count is conducted. Habitat data include number of trees, tree sizes, canopy height, canopy cover, occurrence of streams, etc. Habitat data are collected in a defined radius around the point-count. Observers can estimate this radius using the same bulls-eye “map” described above. The radius usually ranges between 10 and 50 m. From our experience in Bornean plantations, a radius of 20 m provides the maximum area for which a complete census of habitat variables can be conducted accurately in a relatively short amount of time. The choice of habitat variables varies from study to study. A list of variables commonly used in bird surveys is provided in Appendix 3, and a sample habitat datasheet in Appendix 4. Bird and habitat survey data collected in this manner can be analyzed using a variety of tests that focus on community structure or population density. Some of the most basic summary values, including “species richness” (number of species) and “species diversity” (number of species weighted for abundance of individuals within each species) may be computed from any number of programs, including PC-ord (McCune & Mefford, 1999). These summary values provide a basic comparison among plots. Some useful community analyses include species-area or speciessample curves, community similarity, and indicator species analysis. We use PC-ord to conduct these analyses as well. Species-area curves are useful for assessing sampling effort and species richness. They depict how many new species are added to the community list with each new survey sample (Appendix 5); each survey sample represents an increase in sampling area. At some point in any community, conducting additional surveys will not add many new species, and the species-area curve flattens out. In species-rich communities (e.g., primary rainforest), the number of samples required before the curve flattens is very high. In species-poor communities, the number of samples required is low. This suggests that sampling effort should be greater in species-rich habitats than in species-poor habitats, and survey design should reflect this difference. Community similarity is a measure that compares composition and relative abundance of species among communities (e.g., similarity in different aged stands of plantation trees, or between plantation and natural forest). This analysis uses a method called a MultiResponse Permutation Procedure (available in PCord), which is a non-parametric method, similar to an ANOVA, for testing difference among communities. This method provides more information on communities than species richness or diversity indices in that it provides an assessment of overlap and uniqueness of communities. It includes bird survey and habitat data in its comparisons. Indicator species analysis can be a powerful tool in designing focused surveys. Indicator species are species that are indicative of a particular habitat based on their presence and abundance in that habitat compared to others. We determine indicator species in PC-ord which uses the method of Dufrene and Legendre (1997). This method calculates the relative abundance of each species in the dataset across forest types. This value is then tested for significance using a Monte Carlo technique. This method differs from more traditional assignments of indicator species (according simply to rareness) in that it requires a systematic survey design and multiple detections of a species in at least one forest type. Rare species are not likely to be assigned as indicator species because rare species, by definition, are unlikely to be observed many times (if at all) during a survey. The power of this analysis is that indicator species are determined statistically to be more abundant in the forest types to which they a re assigned. Another benefit is that the indicator species assigned using this approach are common enough to be surveyed and monitored over time using straightforward methods. A third benefit of this method is that paraornithologists can be trained to recognize and collect information on indicator species in a short period of time, and they can use this skill to conduct effective surveys that also do not require much time in the field (e.g. one month). We recommend that information gathered on indicator species in different forest types be used in conjunction with comprehensive species lists (which will document rare species occurrence) that are updated every one to three years. Specific example From 19 July to 12 August 2006 we conducted 640 point-counts along 32 transects in five forest types at Grand Perfect: 2-year Acacia mangium (2yAm), 5yAm, 7yAm, secondary forest in the buffer zone, and peat swamp/kerangas forest in Binyo Conservation Reserve. Each transect was placed randomly within 137 Alison Styring, Frederick H. Sheldon, Roslina Ragai and Joanes Unggang each forest type and consisted of 20 point-count stations spaced 50 meters apart. At each point-count station, we conducted a single three minute count using distance sampling methods described above (Buckland et al., 2001). All birds detected during each survey were recorded and their distance from the observer estimated using laser rangefinders. Species richness, species density, area curves, community similarity, and indicator species were determined using the program PC-ord (McCune & Mefford, 1999). The program Distance (Thomas et al., 2003) was used to determine optimal sampling design and sampling effort for detecting focal (indicator) species. DISCUSSION From our preliminary study at Grand Perfect, we were able to determine minimal requirements for future survey work in the plantation. Our analyses of survey methods for Grand Perfect indicated that, across habitats and species, detections of indicator species dropped significantly around 100 m, indicating that the optimal survey distance for these species is 200 m between points. The larger the distance between points, the fewer the surveys that can be conducted during peak birding times (first light to 10 am). However, a significant amount of time is spent at each point (in our experience, at least five to 10 minutes) and the time saved by reducing points means that observers can travel longer distances along a transect. We estimate that observers could conduct at least six point counts (traveling 1.6 km) on a transect survey. Increasing the point-count duration (by, for example, 3 to 5 minutes) would also increase the sample size of observations, but increasing the count time increases bias in the estimates. Thus, traveling longer distances would be preferable to increasing count duration. Thirty point counts per forest type is generally considered the minimum sample size for accurate estimates, but our analysis suggests that larger sample sizes are needed. Optimally, 60 point-counts should be conducted in each habitat type. This equals 10 persondays per habitat, which could be completed in 3–5 days by 2–3 trained observers. Our analyses resulted in a list of 17 indicator species across forest types (Appendix 6). Four of the species listed as indicators in Binyo reserve (Aegithina tiphia, Anthreptes singalensis, Lonchuåra fuscans, and Dicaeum cruentatum) were found almost exclusively in kerangas. Because kerangas is a unique habitat and substantially different in structure from most of the forests in the buffer zone, we excluded these four species from further analyses. Alcedo meninting was another indicator for Binyo, but because it was found near relatively large waterways (in Binyo and at other sites with similar-sized waterways), it was also excluded from further analyses. The remaining species fell into the following taxonomic groups: cuckoos, trogons, barbets, bulbuls, babblers, monarch flycatchers, tailorbirds, and spiderhunters. Of these groups, barbets, trogons, and monarchs were considerably more significant in native forest and older Acacia mangium stands. The remaining groups were comprised of species that replaced one another across forest age (e.g. the tailorbird example stated earlier in this paper). We recommend that trained observers focus on the list of species included in Appendix 7. This list is composed of species determined to be indicators using the method of Dufrene and Legendre (1997) plus some closely related species that were found to be more common in certain forest types than others, but barely missed the 0.05 significance-level cutoff established for true indicator species. An increased sampling effort will likely establish these species as indicators. Paraornithologists should be trained to recognize the songs and calls, and diagnostic field marks of the 33 species listed as focal species for surveys. These observers should then be trained to conduct pointcounts using distance-sampling methods, including the use of laser rangefinders and GPS. These paraornithologists can then monitor key bird populations in plantation stands as they age and compare these populations to those found in the buffer zone and reserves. One way to aid the process of training paraornithologists to identify these species would be compile a song recording for the focal species. In planning surveys, the following observations are important. (1) More surveys are better than repeated surveys; i.e., it is better to survey a larger area (i.e. to conduct a new survey each day during a survey period and for each observer to conduct surveys independently of one another) than to repeatedly cover the same transect. (2) Transects should be established using a random start point and randomly chosen direction. Once the start point is established, the survey should follow as straight a line as possible. This ensures representative coverage of the habitat. (3) To account for variation in habitats across the entire plantation, surveys in plantation forest should be conducted in more than one compartment per age grouping. Preferably, compartments and surveys should be as far apart as possible. (4) As many habitat features as possible within a forest type should be 138 Determining the Diversity of Birds in Bornean Tree Plantations covered in surveys, including streams and roads. Species present and their detectability along streams, logging roads and well-established trails are dramatically different from those found in interior habitat. Thus, it is unwise to have a transect follow a stream, road, or established trail. However, it is desirable to have transects cross streams, roads, and trails in a random fashion to ensure that such features are proportionately represented in surveys of each forest type. (5) The ideal method for transect establishment would to use GIS randomly to select transects. Waypoints on those random transects could then be uploaded to a GPS unit, and observers could navigate to those points. Finally, (6) although we censused three ages of Acacia, if time or resources are limited, it is reasonable to census only two. Young stands should be censused within three years of planting (but at least 1.5 years after planting). Older stands should be censused within two years of the intended harvest. Dufrene, M. and Legendre, P. 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs 67: 345–366. Janzen, D.H. 1993. Taxonomy: Universal and essential infrastructure for development and management of tropical wildland biodiversity. Pp. 100–113. In: O.T. Sandlund and P.J. Schei (eds.). Proceedings of the NNorway/UNEP Expert Conference on Biodiversity. Directorate for Nature Management and Norwegian Institute for Nature Research, Trondheim, Norway. Lambert, F.R. 1992. The consequences of selective logging for Borneo lowland forest birds. Philosophical Transactions of the Royal Society of London B Biological Sciences 335: 443–457. McCune, B. and Mefford, M.J. 1999. PC-Ord: Multivariate Analysis of Ecological Data. Version 4.02. MjM Software, Gleneden Beach, Oregon. ACKNOWLEDGEMENTS Mitra, S.S. and Sheldon, F.H.. 1994. Use of an exotic tree plantation by Bornean lowland forest birds. We thank the staff of Grand Perfect Sdn Bhd and Auk 110: 529–40. Sabah Softwoods Sdn Bhd for their extensive logistical support of our research. We owe a particular debt to Mohd. Hatta Jaafar, Allison Kabi, Mansuit Styring, A.R. and Ickes, K. 2001. Woodpecker abundance in a logged (40 years ago) vs. unlogged Gamallang, and Elizabeth Bacamenta for help at lowland dipterocarp forest in Peninsular Malaysia. Sabah Softwoods, and Rob Stuebing, Nyegang Journal of Tropical Ecology 17: 261–268. Megom, and Latiffah Waynie, Stephven Stone, Henry Nyegang, Last Gundie, Kelvin Bryan, and Li Joseph Styring, A.R. and Zakaria, M. 2004a. Foraging for help at Grand Perfect. ecology of woodpeckers in lowland Malaysian rain forests. Journal of Tropical Ecology 20: LITERATURE CITED 487–494. Basset, Y., V. Novotny, S.E. Miller and R. Pyle. 2000. Quantifying biodiversity: Experience with Styring, A.R. and Zakaria, M. 2004b. Effects of logging on woodpeckers in a Malaysian rain parataxonomists and digital photography in Papua forest: the relationship between resource New Guinea and Guyana. BioScience 50: availability and woodpecker abundance. Journal 899–908. of Tropical Ecology 20: 495–504. Buckland, S.T., Anderson, D.R., Burnham, K.P., Laake, J.L., Borchers, D.L. and Thomas, L. 2001. Thomas, L., Laake, J.L., Strindberg, S., Marques, F.F.C., Buckland, S.T., Borchers, D.L., Anderson, Introduction to Distance Sampling: Estimating D.R., Burnham, K.P., Hedley, S.L., Pollard, J.H., Abundance of Biological Populations. Oxford and Bishop, J.R.B. 2003. Distance 4.1. Release 2. University Press. Oxford. 432 pp. Research Unit for Wildlife Population Assessment, University of St. Andrews, UK. Diamond, A.W., K.D. Bishop and S. Van Balen. 1987. Bird survival in an isolated Javan woodland: Island or mirror. Conservation Biology 1: Wong, M. 1986. Trophic organization of understory birds in a Malaysian dipterocarp forest. Auk 103: 132–142. 100–116. 139 Alison Styring, Frederick H. Sheldon, Roslina Ragai and Joanes Unggang Appendix 1. Bulls-eye map for surveys. Date:_______Time:_____ Name:__________ Point number:_____ Duration:_______ Instructions 1. Measure land marks that represent near (inner circle) and far (outer circle) distances from your point.You may also indicate other relevant landmarks on the sheet. 2. List all species seen and heard during the count on this sheet 3. Indicate if the alert cue was visual (V) or aural (A) 4. Measure the strait-line radial distances of the species with rangefinders 5. If an individual is too far away to estimate, give your best estimate and indicate this with an “E” beside the species name 140 Appendix 2. Sample point-count datasheet. Determining the Diversity of Birds in Bornean Tree Plantations 141 Alison Styring, Frederick H. Sheldon, Roslina Ragai and Joanes Unggang Appendix 3. Habitat variables. Point number Survey start time (three minute duration) GPS coordinates (in UTM, zone 50, WGS84) Easting Northing Elevation Slope: 4 categories—(1) 0–5%, (2) 5–15%, (3) 15–25%, (4) >25% Habitat 1 - description Habitat 1 % coverage Habitat 2 - description Habitat 2 %coverage Standing water (% coverage) Stream width (m) Adjacent land use (100 m radius) Logging road/treefall gap Number of forest layers (max 4) Canopy height (m) Canopy % coverage (rounded to nearest 10%) Secondary canopy height (m) Secondary canopy (% coverage) Shrub height (to nearest 0.5 m) Shrub (% coverage) Ground cover height (to nearest 10 cm) Ground cover (% coverage) Number of woody stems in a 5 m2 extra-small (e.g. <5 cm) small (e.g.5–10cm) medium (e.g.10–25 cm) large (e.g. 25–40cm) extra-large (e.g.>40 cm) Weather 142 Appendix 4. Sample habitat datasheet. Determining the Diversity of Birds in Bornean Tree Plantations 143 Alison Styring, Frederick H. Sheldon, Roslina Ragai and Joanes Unggang 144 Determining the Diversity of Birds in Bornean Tree Plantations Appendix 5. Species-Area curves for Grand Perfect. 145 Alison Styring, Frederick H. Sheldon, Roslina Ragai and Joanes Unggang Appendix 6. Indicator species for 2-year, 5-year, and 7-year Acacia mangium, conservation buffer, and Binyo Conservation Research Area at Grand Perfect Plantation. An “X” indicates that the species was determined to be an indicator for that specific forest type. Indicator species Cacomantis merulinus Harpactes kasumba Alcedo meninting Megalaima rafflesii Aegithina tiphia Pycnonotus erythropthalmus Macronous gularis Pellorneum capistratum Stachyris maculata Malacopteron magnum orthotomus sericeus orthotomus ruficeps Terpsiphone paradisi Dicaeum cruentatum Anthreptes singalensis Arachnothera longirostra Lonchura fuscans P-value 0.01 0.01 0.04 0.01 0.02 0.03 0.02 0.04 0.01 0.01 0.01 0.01 0.03 0.01 0.02 0.01 0.01 2-y AM 5-y AM 7-y AM Buffer Binyo C.R. X X X X X X X X X X X X X X X X X 146 Determining the Diversity of Birds in Bornean Tree Plantations Appendix 7. Recommended focal species for surveys at Grand Perfect Plantation. • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • Cacomantis merulinus Cacomantis sonnerati Harpactes kasumba Harpactes diardii Harpactes duvauceli Megalaima rafflesii Megalaima australis Megalaima chrysopogon Megalaima mystacophanos Megalaima henricii Pycnonotus erythropthalmus Pycnonotus simplex Pycnonotus brunneus Pycnonotus atriceps Pellorneum capistratum Macronous gularis Macronous ptilosus Stachyris maculata Stachyris erythroptera Stachyris nigricollis Stachyris rufifrons Malacopteron magnum Malacopteron cinereum Malacopteron magnirostre Malacopteron affine Orthotomus sericeus Orthotomus ruficeps Orthotomus atrogularis Terpsiphone paradisi Arachnothera longirostra Arachnothera robusta Arachnothera flavigaster Arachnothera crassirostris 147