Bird diversity in northern Myanmar and conservation implications

We conducted four bird biodiversity surveys in the Putao area of northern Myanmar from 2015 to 2017. Combined with anecdotal information collected between 2012 and 2015, we recorded 319 bird species, including two species (Arborophila mandellii and Lanius sphenocercus) previously unrecorded in Myanmar. Bulbuls (Pycnonotidae), babblers (Timaliidae), pigeons and doves (Columbidae), and pheasants and partridges (Phasianidae) were the most abundant groups of birds recorded. Species richness below 1 500 m a. s. l. was higher than species richness at higher elevations. Our results suggest that the current protected areas in this region should be expanded to lower elevations to cover critical conservation gaps.


INTRODUCTION
The northern part of Myanmar's Kachin State borders both China and India and lies in the Indo-Burma conservation hotspot (Myers et al., 2000). Although this area harbors rich biodiversity, conservation-related research in this region is lacking due to its steep topology, difficult transportation, and variable climate (Rao et al, 2010;Rappole et al., 2011). Since 1990, one new species of mammal (Muntiacus putaoensis) (Rabinowitz et al., 1999) and several new subspecies of birds, such as Tesia olivea chiangmaiensis (Renner et al., 2008), have been described in this region. This indicates that there are still opportunities for new discovery and exploration and that more information is needed for future conservation plans.
Since the 1990s, several bird surveys had been carried out in the Putao area (Rappole et al, 2011). Under the leadership of the Nature and Wildlife Conservation Division (NWCD) of the Myanmar Forestry Ministry, two expeditions were launched in 1997-1998 (Aung & Oo, 1999) and 2001-2009(Rappole et al., 2011, providing the most detailed inventory of local avian diversity thus far. 1 Between December 2015 and May 2017, the Southeast Asia Biodiversity Research Institute, Chinese Academy of Sciences (CAS-SEABRI), Forest Research Institute (FRI) of Myanmar, Hponkan Razi Wildlife Sanctuary (HPWS), and Hkakabo Razi National Park (HKNP) jointly conducted four general biodiversity surveys in the Putao area of the northwest part of Kachin State, Myanmar. The survey region covered part of HPWS and the surrounding areas south and northeast of Putao ( Figure 1). These surveys were conducted to acquire basic biodiversity distribution data for future conservation policy making.
In this paper, we summarize the results of our four surveys as well as episodic observations of a local officer (Myint Kyaw, M.K.). The present research aimed to: (1) update the avian inventory of the study area; and (2) identify survey and conservation gaps in this region.

Study area
The survey area included the southern part of HPWS, one village south of Putao town, and areas northeast of Putao surrounding HKNP (Figure 1). The total survey area was around 5 000 km 2 . Elevation in the area varies from 400 m a.s.l. at Putao to 5 881 m a.s.l. at the Hkakabo Razi Snow Mountain peak (Renner et al., 2007). Vegetation includes lowland tropical rainforest dominated by Dipterocarpaceae between 400 m a.s.l. and 600 m a.s.l., tropical seasonal rainforest dominated by Terminalia myriocarpa and Dysoxylum sp. from 600 m a.s.l. to 1 200 m a.s.l., mountain tropical rainforest between 1 200 m a.s.l. and 2 000 m a.s.l., subtropical evergreen broad-leaved forest dominated by Lauraceae and Fagaceae between 2 000 m a.s.l. and 2 600 m a.s.l., and bamboo-rhododendron habitats above 3 000 m a.s.l. (Renner et al., 2007;Yun-Hong Tan, personal communication). The temperature varies from 2 °C to 36 °C. Annual precipitation ranges from 3 800-5 100 mm, with a cooler dry season from October to February and a rainy season from March to September (Robinowitz et al., 1999).

Bird survey
The surveyed habitats include cropland, small areas of riparian wetland (less than 10 count points), and the forest vegetation types mentioned above. Survey locations ranged from 400 m a.s.l. to 3 200 m a.s.l.. We conducted field surveys from 29 November 2015 to 4 January 2016; 23 April to 21 May 2016; 26 November 2016 to 5 January 2017; and 28 April to 26 May 2017.
Point counts and camera traps were combined to obtain bird diversity information. Points were situated along trails from 400 m a.s.l. to 3 200 m a.s.l.; each point was located at least 300 m from other points to ensure quasi-independence (Ralph et al., 1995). We stayed at each point for 12 min, recorded all birds seen or heard within a 100-m radius. We used camera traps to detect ground-dwelling birds and used binoculars to detect all other birds. In total, we established 53 camera stations ranging from 700 m a.s.l. to 3 200 m a.s.l.. The distances between any two camera stations always exceeded 500 m. The cameras were installed in the field between December 2015 and May 2016. The distribution of the point counts and camera traps were mapped out to show the sample effort at different elevations. In addition to systematic data obtained from point counts and camera traps, species observed during field trips were also recorded. The second author M.K., who works at HPWS and HKNP, is an experienced bird-watcher and collected anecdotal information and data through photography and direct observations between 2012 and 2015.

RESULTS
During the four surveys, we collected information from 304 count points. Of the 53 cameras installed, one was stolen and one was broken. The remaining 51 cameras accumulated more than 7 500 working days of images.
In total, 319 species were recorded, representing 72 families and 18 orders (Table 1), with one species listed as critically endangered, five species listed as vulnerable, and 13 species listed as near threatened in the IUCN Red List (IUCN, 2016). The percentage of birds of each family is shown in Figure 2. In terms of number of individuals recorded during the surveys, the most abundant families were Pycnonotidae (989 individuals), Leiothrichidae (850 individuals), and Columbidae (608 individuals). Several families recorded, such as Alaudidae, Artamidae, Indicatoridae, Psittacidae, Troglodytidae, and Upupidae, contained only one identified individual. The number of different species at different elevations is shown in Figure 3. The distributions of the point counts and camera traps at different elevations are shown in Figure 4. We observed 3 719 (82%) individuals from 215 (67%) species under 1 500 m a.s.l.. Sampling effort was biased towards elevations below 1 500 m a.s.l..   ). It was identified by its large body size, black facial mask, grey nape and upperparts, white scapular, and white spot on the wings.

DISCUSSION
Because our data came from direct observations and camera trap photos, some species could be under sampled, particularly those living in shrub and canopy habitats. Our sampling effort was more intensive in lower elevations, partly because a larger proportion of the survey area occurred at lower elevations and partly because of easier transportation in the lowland plain. The overrepresentation of lower elevation habitats in our study could explain the low frequency of many bird families in our surveys. In the future, surveys should be conducted at higher elevations, especially in the mountains above 3 000 m a.s.l.. Although our sampling effort and checklist is incomplete, the present work provides preliminary data for local diversity research and conservation projects.  Rappole et al. (2011) conducted five bird surveys in a 10 000 km 2 area in the Hkakabo Razi region, mainly covering the east. While the present study area also lies within the same region, we focused on the western part of Putao. Furthermore, Rappole et al. (2011) set mist nets from 500-3 000 m a.s.l., and recorded 413 species, of which 159 were not observed during our surveys. Of the 319 species recorded in the present research, 65 were not recorded by Rappole et al. (2011). The greatest difference in species occurred in the families Muscicapidae (n=7) and Accipitridae (n=6). This disparity between studies is likely explained by the different study areas sampled as well as the differences in methodology. Our methodology was more suited for detecting certain species. For example, birds of prey (Accipitridae) can be more easily observed during point counts, and some cryptic species can be captured by camera trap. In our study, one Eurasian Woodcock (Scolopax rusticola) was captured by camera trap at 1 190 m a.s.l.. We believe future exploration of new survey routes and a combination of different survey methods will lead to the discovery of additional species not previously recorded here.
The Chestnut-breasted Partridge was identified by its rufescent crown, white gorget, and chestnut breast. It was previously only known from Bhutan, Southeast Tibet, China, and Northeast India (McGowan et al., 2017). The Chinese Grey Shrike was previously recorded in eastern Mongolia to southeast Russia, northern China, and North and South Korean, also known to migrate to east and southeast China and Korea (Yosef & International Shrike Working Group, 2017). The finding of these species formerly unrecorded in the area not only increase two new species of birds for Myanmar, but also reiterates the high biodiversity in this region and highlights the importance of long-term field observations.

Figure 7 Photograph of Chinese Grey Shrike (Lanius sphenocercus) (photo by Myint Kyaw)
Based on our observations, the forest is in good condition. The Leiothrichidae, Columbidae, and Phasianidae species are forest-adapted, and their high abundance shows adequate food and resources for breeding in the local forest. Hornbills were observed almost every day around the camp sites at 1000-1 500 m a.s.l. during winter, and gibbons (Hoolock spp.) were often heard from the camp sites below 2 000 m a.s.l.. The most abundant species observed belonged to Pycnonotidae, Leiothrichidae, Columbidae, and Phasianidae. The bulbuls (Pycnonotidae) are common birds living in tropical and subtropical areas, which can adapt to diverse habitats (Fishpool & Tobias, 2017).
In addition to the forest, the local cropland also provides important habitat for many migrating species. Farmers in Putao plant large areas of rice in June-July and harvest it in October-November. The land is barren in other months of the year. These fallow lands provide important habitat for migrating bird species. M.K. observed more than 10 000 common cranes (Grus grus) migrating through Putao in March 2016, where they forage in the fallow rice paddies and wetlands by the Malika riverside. Figure 3 shows that both abundance and species richness were higher below 1 500 m a.s.l.. Species of high conservation value, (e.g., critically-endangered White-bellied Heron (Ardea insignis), vulnerable Rufous-necked Hornbill (Aceros nipalensis), and near-threatened Lesser Fish Eagle (Haliaeetus humilis)) were all recorded under 1 500 m a.s.l.. We also observed some birds with narrow ranges such as the Snowy-throated Babbler (Stachyris oglei) at lower elevations. Current protected area boundaries are set too high in minimal elevation (1 000 m a.s.l. for HPWS and 900 m a.s.l. for HKNP) and should be lowered to accommodate greater biodiversity, a conclusion also drawn by former researchers (Rabinowitz et al., 1999;Renner et al., 2007). The Burmese government currently plans to extend these two protected areas to incorporate lower elevations, and is applying for World Heritage Site status for these two regions. This is an important decision with long-term benefits for local conservation. The HPWS forest connects to the Hukawng Valley extension in the west. Extending the current HPWS and HKNP borders will ensure the protection of the largest forest complex in northern Myanmar and will maintain intact habitat for important wildlife.