Japanese Yellow Bunting merits concern regarding status with low numbers recorded and vulnerability to trapping

After I saw two Japanese Yellow Buntings (aka simply Yellow Buntings Emberiza sulphurata) in Hong Kong this spring, I happened to notice that Birdlife has changed the species’ status from Vulnerable – as it was from 1994 to 2016, to Least Concern in 2022.

Japanese Yellow Bunting

This surprised me: after all, this species has a restricted range, breeding in Japan and wintering mainly in the Philippines (and Taiwan), and perhaps a low population. In Hong Kong it’s a scarce migrant, mainly in spring – probably as northbound birds from Luzon are deflected west in easterlies winds.

So I had a look at the evidence on the Birdlife pages about this species; and it turned out there is huge uncertainty in the population – with perhaps 2500-9999 mature individuals, maybe as many as 100,000 breeding pairs in the early 2000s. Yet it appears the estimates are based on extrapolations from fairly small numbers found in breeding territories, with no high counts anywhere that I could find – so most of them are imaginary!

[While there is a tendency to refer to this species as “Yellow Bunting”, this name has also applied to Yellowhammer; and I think might be confused by some with Yellow-breasted Bunting. Hence, I think Japanese Yellow Bunting is a better, name, with no chance for confusion, and reflecting it only breeding in Japan.]

And even if you skim this info, which I’ve added to at times, maybe have a look at this – from eBird, showing all records of this bunting from throughout its range. Sorted by numbers observed; with only one triple figure count, vast majority of records involving less than 20. Also repeated later in this article.

Yellow Bunting Assessment in 2001: Vulnerable

There is a lengthy account of [Japanese] Yellow Bunting in Threatened Birds of Asia: The BirdLife International Red Data Book, published in 2001. This account is available online as a pdf file; it begins:

This bunting qualifies as Vulnerable because it has a small and declining population, probably through a combination of habitat loss, pesticide use and hunting throughout its range.

The account has an extensive collection of records from throughout the species’ range; even includes one I saw on island where I live, Cheung Chau (an April bird, on a banana leaf rather than standard habitat at/beside lowland grassland/ wetland). Reading through fairly quickly, there are plenty of records, but mostly single figures, often single birds; highest count I notice is 46.

There are historical reports from early last century of the species being “abundant” on the slopes of Mount Fuji, and in perhaps one province in north Luzon. However, it appears there was a substantial decline during last century; at the time the account was written, it was unclear whether the decline was continuing;

However, given the various threats that it faces on migration and in winter, and the general decline in the numbers of small passerines that has been noted in Japan, Korea and mainland China (see Threats), it is here assumed that it may well still be in decline.

Stable population?

Evidence for a stable population at present is largely from breeding area records, chiefly from the atlas of Japan’s breeding birds. Yet in this, (Japanese) Yellow Buntings have most recently been found in 97 of 2344 sites; and confirmed breeding in only 8 quadrats, with possible breeding in another 118 quadrats (see info via this link):

Japanese Yellow Bunting in atlas of Japan’s breeding birds

There is also mention of extrapolating from a survey of the species on breeding grounds; as if could multiply density of 0.5 per hectare found in this, by area of similar habitat. Yet extrapolating from small number to large/very large ones [maybe tens of birds to thousands, in this case] is not such a robust thing to do. And difficulty surely compounded as Japanese Yellow Bunting has a patchy distribution even within its breeding range: see below.

The Birdlife info also refers to a paper on buntings in Korea, which has numbers of ringed birds, with 127 Yellow Buntings ringed from 1990s-2010s, compared to 11 from 1950s to 1980s: again, not high numbers for basing a conclusion on.

Japanese Yellow Bunting scarcer in Hong Kong

While Hong Kong is likely on the periphery of the species range, substantially higher numbers have been recorded here than ringed in Korea, with up to 116 in a spring prior to 1996; but since then, only single figures in any spring, despite increased observer coverage.

Hence, it appears there has been a substantial decline in Japanese Yellow Buntings occurring in Hong Kong.

This appears to be the only place with such detailed data over the years, though perhaps there is more from other migration sites such as Japanese islands, Taiwan: if so, would be interesting. Otherwise, currently like having a very poor picture of an iceberg [species in its main range], while monitoring of tip of iceberg suggests all is not well.

Flocks north Philippines and while se China is big the coastal area is slender

I’ve started a Facebook thread about this bunting, eliciting a few comments.

Among them:

They mostly winter in the northern Philippines in agriculture that is rarely visited by birders, I’ve seen flocks of over 100 when I’ve looked though.

Robert Hutchinson

Yet also, no surveys conducted over the years; no trends. But at least substantial numbers.

Also, after I’d noted there is a lack of habitat for this species near Hong Kong:

there’s massive amounts [of habitat] for several other species of wintering and migrant buntings that have similar habitat niches.

James Eaton

Well, I’m not at all sure where these massive amounts of bunting habitat are in the species’ migratory range in south and southeast China – not in the Pearl River Delta area near Hong Kong, nor elsewhere in the southeast, especially given migrants evidently favour lowlands, especially grasslands in or by wetlands [as Hong Kong].

Southeast China – mostly hilly, with slender lowlands along coast; surely not massive amounts of habitat for Japanese Yellow Bunting
An example of coastal lowland in the area – Xiamen and environs; looks heavily developed, not so much suitable area for buntings perhaps

As eBird records indicate, migrating Japanese Yellow Buntings indeed appear mostly restricted to this coastal strip, and to nearby islands (as I write, 1728 of the 2499 records are from Taiwan):

Japanese Yellow Bunting records on eBird

I contacted Beijing based ornithologist Xu Shi re Japanese Yellow Bunting status becoming Least Concern; he replied:

That’s surprising to me as well.  Japanese Yellow Bunting records in Mainland China are really sparse—mainly from eastern coastal provinces, during migration seasons in similar habitats you just described.  I’m not quite familiar with its numbers so it is hard to estimate a trend. 

Xu suggested I contact Wieland Helm, who is studying Asian buntings. I did so, and he replied:

thank you very much for this interesting information. Yes, many East Asian species are listed as “Least concern” simply because of the lack of monitoring data. Its a shame. I suggest you send a link of your blog post to BirdLife again, maybe they will consider the status of the species again in the next round of revisions.

Perhaps small population, restricted range, possible trapping

As noted on the species’ Birdlife page, several bunting species in the region are in decline, with an especially severe crash in numbers for Yellow-breasted Bunting – which is now rated Critically Endangered.

Trapping is a key reason for the precipitous recent decline in Yellow-breasted Bunting; mainly for consumption in south China. Trapping has been significant for Japanese Yellow Bunting in the past; as noted by Deguchi et al:

In the early decades of the 20th century, as many as 30,000 or more Yel- low Buntings were caught in a single year (Uchida 1937), a number that greatly exceeds the current esti- mate of the entire population of the Yellow Bunting

Temperature and snow depth explain the breeding distribution pattern of Yellow Bunting Emberiza sulphurata

Does it affect Japanese Yellow Bunting? Perhaps no one knows, but there evidently is bird trapping in se China, which can be substantial – a Google search reveals:

On April 10th [2020], the China Biodiversity Conservation and Green Development Foundation (CBCGDF) research team tracked and reported that a kilometer-long birding net was discovered in a 200-mu fishpond in Changting County, Longyan City, Fujian Province. 

The Kilometer-long Bird Catching Net Has Been Quickly Cleared

Wu, 39, … was captured carrying two large bags filled with dead birds in the Chinese province of Zhejiang on Nov. 7 [2022].

Man in China arrested for killing 1,000 birds ‘for hotpot’

These birds included 948 “sparrows”; here perhaps a blanket term for sparrow/ finch/ bunting type birds.

the nets are going up faster than they can be taken down. The volunteers have found that online shopping sites have spurred the trade in captured birds. They complain that Taobao, China’s largest online retailer which is owned by Alibaba, has made it easier and cheaper for hunters to acquire tools, meaning disaster for migratory birds and other animals.

A search on Taobao for “bird nets” brings up 5,000 suggested purchases, including one net that is 5-metres high and 30-metres wide for only 30 yuan (US$5). Cheap prices and quick delivery makes acquiring nets simple.

Migratory birds in peril from trappers

The latter begins by considering the coast east of Beijing – evidently a key area for trapping Yellow-breasted Buntings during their migration; but it also indicates trapping is all too common, helped by nets and other paraphernalia being cheap and readily available.

Surely it would be wise to try and ascertain if any Japanese Yellow Buntings might be caught this way – as trapping in their limited migratory range could be a major issue for the species. Indeed, I don’t think it would be a great stretch to suggest they might be trapped deliberately, for sale as supposedly the valuable and now rarer “Yellow-breasted Buntings”.

Update: just seen link to forum discussion from January 2022, Yellow Bunting (Emberiza sulphurata): Revise global status? While this led to a revision, from Vulnerable to Least Concern, based on info as I’ve noted above – especially breeding atlas results – I notice a comment:

Some trapping and trade is recorded on Taiwan and moderate domestic trade is recorded in the Songbirds in Trade Database (SiTDB) suggesting the species might be targeted. 

It seems this comment was ignored; I didn’t see it referred to in the Birdlife info on the red list status, published after this forum post/discussion. But also further evidence the potential for trapping having an impact should be considered.

Hence, while there may be no need for urgent alarm bells for Japanese Yellow Bunting, perhaps cagetorising as Least Concern is just too relaxed.

A response of sorts from Birdlife International

Of course, it would seem obvious to check with Birdlife International about this species’ status. I emailed, and received this response, which I didn’t find helpful:

Dear Martin,

Thank you for your email, it is nice to hear of your care an interest in birds.

Our Hong Kong partner Hong Kong Birdwatching Society (HKBWS) is the best resource for information on local species populations.

They can be contacted at [email protected]

Kind regards


Well, I happen to be a member of HKBWS, and have some familiarity with local species; plus was writing with concern about Japanese Yellow Bunting’s world population. Ah well, seems Claudia is a gatekeeper for Birdlife’s main email address, so my enquiry fell flat, concerns not addressed.

Further Correspondence with Birdlife International

Happily, I later received a far more nuanced response from Birdlife, from Senior Red List Officer Rob Martin:

The forum consultation for Yellow Bunting closed last year and the reassessment at Least Concern was published in December. That trapping has been flagged and may continue to affect individuals of the species is noted in the threats section, but the data from the Japanese breeding bird atlas strongly suggests that the breeding population is not declining. With this data pertaining to the breeding population, it is very difficult to make a case that threats elsewhere in the range are causing a significant population decline, given that is not thought to breed elsewhere. A couple of hypothesis spring to mind as to why declines in a species may be occurring along migratory routes but unless there is an alternative unsurveyed breeding population or the Japanese Bird Atlas data are wrong, then neither suggest that there are likely to be declines sufficient to impact the population size.

One hypothesis is that there is a decline, but that there is a surplus of individuals that do not appear in known breeding areas.

Second, migratory routes have altered. 

The first appears unlikely for a bunting. The second merits investigation. We are aware of research demonstrating dramatic shifts in breeding range for species in China, which suggests that conditions favouring certain migratory routes may also change. Obviously, extreme trapping is itself a strong selection pressure favouring individuals that do not use that route: historical rapid declines are cited for the species and these may have shaped current usage of the migratory pathways available to the species. Or they might be short-stopping, or undertaking one-stop strategies versus coast-hopping.

With the data available, the assessment at Least Concern is the only available against the IUCN Red List Categories and Criteria.

The species will be reassessed before 2030, and all new information will be considered for that assessment.

Rob Martin

Dear Rob:

More a response as I’d hoped for; been a few years since I emailed Mike Crosby, wondering about status of grass-warbler/ grassbird, as it seemed to me there were few records in vicinity of Hong Kong. That was before the grassbird here became split to a separate species…

As to [Japanese] Yellow Bunting, my main concern is – as it were – Least Concern may mean too much relaxation, when it merits attention.

I have read through the info on the Birdlife pages about Japanese Yellow Bunting [longer winded name, but clearly not Yellowhammer!]; also today seen account from Threatened Birds of Asia.

Really, little has changed since the Threatened Birds of Asia account (Collar 2001).

Yes, I see Japanese Bird Atlas, but not so many quadrats, especially if look at breeding. And how many per quadrat? [I asked a friend who does a lot of birding in Japan; he told me common in parts of Honshu, but still doesn’t know number].

Mark Brazil of course a massive figure in Japan/e Asian birding; yet if the population is large based on extrapolations, where are they? 

Where are the records of tens and more of Yellow Buntings?

I’ve posted on FB; Robert Hutchison noted he could see flocks of 100 in n Luzon. 

But otherwise, largest count I saw in Collar 2001 was 46; maybe I overlooked higher. And mostly far less; even my single record from island I live on is included.

I believe the HK info deserves consideration; maybe best proxy there is for Luzon population trend:

//In contrast, from the mid-1980s to 1998 small parties of birds were not infrequent, and larger numbers occurred during years with influxes, the highest counts being 15 on 11 April 1993 and 17 on 6 April 1996. From 1987 to 1996 there were four years in which the sum of aggregate weekly counts in spring reached double figures or, in one year, a triple figure (116). Since then, however, despite increased observer coverage, especially since 2010, only single figure totals have been recorded. This would seem to indicate that there has been a decline in numbers passing through HK.//

– from the as yet unpublished new Avifauna of Hong Kong, will appear online. 

Geoff Carey sent to me; he too was surprised re change of status, as was Xu Shi in Beijing, who told me records are sparse on se China coast.

Weather seems important for HK records; best if easterlies. Even so, the reduced numbers seem consistent; and habitats here not so much worse, even as nearby Pearl River Delta region more concreted.

So, Hong Kong numbers might indicated a reduction in Luzon population – my guess.

Might be a shift, such as more wintering in Taiwan, with warming?  

Most eBird records are from Taiwan; seems I’m not privy to the numbers recorded at each site though, not easily anyway.  [photos on eBird are of single birds, that I noticed; not flocks… ]

As to its migratory routes otherwise altering, doesn’t seem indicated by records, including the eBird sightings – unless, perhaps, Taiwan now more important.

Trapping is surely the x factor, especially on se China coast, where development has reduced suitable habitat, making birds like this more vulnerable to any trapping.

I did find examples of trapping, via quick google searches. But can be hard to know about trapping, without people in area reporting – cf nets in spoon-billed sandpiper areas sw China found by Jonathan Martinez.

We know, too, re Yellow-breasted Bunting; to me, almost like one minute not much to worry about, and the next, critically endangered. 

Seems there is still demand for ybr bunt; and it’s harder to find and sell, with some action to try and restrict trapping/trade. 

While if anyone wants species to trap and pass off as “rice birds” – well Yellow Bunting would fit this better than many others; just put up some nets along Zhejiang and other coasts, and maybe can profiteer. 

As there is/has been a substantial organised business selling rice birds, not too implausible. [heck, years ago I was told that barbecued “sparrows” on sticks for sale in Beijing were often day old chicks]

So I hope that, as in past terminology, Japanese Yellow Bunting remains among the “Birds to Watch” – not of little/least concern, but very much one to keep an eye on, advocate more attention, more research.


Given the species past as a threatened bird and the commitment of the Japanese government to ongoing biodiversity monitoring I don’t think Yellow Bunting will be ignored to the point that if a decline was suspected in Japan, even before atlas results, this would be publicised and would see a reassessment. It is very much the population trend that is important, as without a decline the Red List only cedes a species may be threatened if there are fewer than a thousand or they are extremely restricted with an potentially catastrophic threat to that restricted area. Hence the uncertainty over the population size is of less concern when breeding atlas data indicates an increase in occupied squares with comparable effort. The population question boils down to how plausible it is that there are only 1,000 mature individuals. With recorded breeding densities of 30-50/km2, it seems not so likely and could be rather a lot per occupied atlas quadrat (paper here: Habitat preferences of breeding Yellow Buntings Emberiza sulphurata in hilly rural areas following rice field abandonment in northern Japan).

Getting a reasonable bound on population size is a crucial next step: it could be fewer than 10,000, in which case any future evidence of a population decline could see it back at Vulnerable.

But we must be guided by the available data and for this species that is evidence of stability or an increase in breeding area.


Just noted to Alex Berryman: since the data is “robust”, and not data deficient, it should be pretty easy to calculate a known world population based on actual records of Japanese Yellow Buntings.

What is this tally?

[if this can’t be calculated, then the data is clearly lacking]

Thanks for the link; quite something this is at least partly a landslide specialist!

Seems an open question as to whether abandoned farmland becoming overgrown can prove an issue.

I also found a paper by Kaneko, with 33 JY Bunting sites found in surveys, at pretty near sea level. Seems to include habitats more as found in Hong Kong, by lowland wetland in at least one site.

Reasonable bounds can be tough. A figure based on actual records would seem a starting point, as informative. [I think of Linnet, say; seen concerns numbers may be falling in UK, but also flock of well over a thousand, via my twitter feed rather than looking into this.] 

Also here, as I also noted, not looking at entire range being some “here be dragons” wilderness; migration routes through areas now with quite a number of birders, bird photographers, including se China [quite a change from when Threatened Birds of Asia published].  “sparse” according to Xu Shi; while Menxiu Tong told me of “lots” seen on a tour to Zhejiang this spring, without following up about what this means.

Taiwan sightings should be informative; I’m about to email ebird folks, see if I’m overlooking something re numbers being available. The eBird photos I saw are of solitary individuals.

Even n Luzon, not much to go on re abundance it seems.


“Robust” and data certainly not deficient; so why no figure for known world population?

Here, referring to back and forth messaging on Facebook thread I started, mentioned above, with the chief responses from Alex Berryman of Birdlife, including:

To approach each of the four Criteria [for the IUCN Red List] in relation to Yellow Bunting:

Criterion A: population reduction. To qualify as NT or VU, a species would have to have declined, or be predicted to decline, by more than 20% or 30% (respectively) over ten years. As the data published in the BirdLife account document adequately, there is no indication this is the case. =Least Concern.

Criterion B: small range + a continuing decline in a set of parameters. This species is not even close to meeting the small range thresholds. = Least Concern.

Criterion C: small population size + a continuing decline. No evidence the population falls below the threshold, and even less evidence for a continuing decline (trend data suggest a stable trend at their most pessimistic). We cannot ignore the Japan trend, which is the most robust available data to us. Is it perfect? No. But knowledge gaps do not preclude assessment and if the situation changes or deteriorates, the species will be re-evaluated. =Least Concern.

Criterion D: very small population size (<1000 mature individuals). No evidence or reason to suspect this is true. =Least Concern.

Alex Berryman

So what is the population? For the most part, guessology; relies on substantial extrapolation. But if it’s so common as to be of little concern, where are they all? 

Collar 2001 list multiple records; but my scanning of this leads to max count of 46, and mostly single figures; also “abundant” only early last century, Mt Fuji slopes and a province in n Luzon, otherwise not so many even then.


Data Deficient certainly doesn’t apply here. 

… By most definitions of extinction risk theory, the species certainly does not have a particularly small range, no (it does not at all approach either of the spatial metric thresholds we use…or any others I can think of). Combining the Atlas data, eBird data, and descriptive accounts (including the one of ‘common’ you just cited) with a quick calculation of the area of habitat using satellite data, it is impossible to reach a conclusion that a population size estimate anything fewer than several thousand would be inappropriately pessimistic. Again, the IUCN Red List is ultimately a metric for *global* extinction risk. If our friends in Japan are saying the species is doing just fine (which they are, we have internal correspondence suggesting this too), then the risk of extinction, as it stands, are minimal. 

Alex Berryman

Given the data is “robust” and not deficient, it should be pretty easy to calculate the known world population of Japanese Yellow Bunting, based on actual records.

What is this tally?


I do not think that would be an easy exercise. I think you are conflating (and possibly confusing) two important principles of data science and derived uncertainty: the issues of accuracy and precision. The breeding data we have from Japan are, indeed, patchy and do not allow us to derive a particularly *precise* global estimate of abundance. But given Deguchi et al. (2017) found breeding densities of 50 birds/km2, and the area of suitable habitat is presumably rather large (an assumption defendable using satellite data *and* the occupancy rates derived from the breeding bird atlas survey), we can have confidence that the BirdLife statement that “the population size…is very likely to exceed thresholds for listing as threatened” is robust and accurate. I.e., we can have can have confidence in the accuracy of a statement, without necessarily needing precise numbers behind it.

Alex Berryman

Thanks again, Alex; well my academic background’s in chemistry inc physical chemistry; I remember a phys chem lecturer talking of “arm waving” when no precision.

When looking at possible changes in species’ occurrences, I’ve also used actual sightings, where possible [hard when get “abundant” or “common”, say].

So to me the data is not robust, nor adequate.

A calculation of known world population based on actual records would be informative – a lower bound for the species. 


Even if some fudging; might include numbers on migration and add these to breeding birds, say; figuring reasonable to assume no overlap.

Then, can come up with a hypothetical figure, based on the satellite data, habitat info [avalanche areas quite a niche! tho I also note Kanecho finding them breeding beside reedbeds – akin to the favoured habitat on migration].

That occupancy rate is 97 out of 2344 atlas survey sites [4%]; occurring in 143 quadrats. 

As I recall, that same phys chem lecturer started with a quote from Ernest Rutherford: “All science is either physics or stamp collecting.”; he also said, “If your experiment needs statistics, you ought to have done a better experiment.”


eBird records mostly just single figures; just 10 of 30 or more

While here, some data from eBird; sorted by numbers observed.

Only one triple figure count; less than 83 records involving double figure counts [some duplicates], inc just 10 of these with 30 or more.

Otherwise, single figure counts.

2030 of 3449 records are of single birds.

[evidently quite some duplicates – looks like at times several observers submit lists for same place, including same birds: so yes, not easy to come up with rough total. For last year, around 1000 if add all sightings, but duplicates comprise quite a proportion of this.]

eBird is perhaps a bit “new-fangled” for trying to assess trends; 

Increased reports of Japanese Yellow Bunting shown by a trend chart (may reflect increased usage of eBird in the species’ range; haven’t checked); yet last year only six of these with double figure counts – n Luzon mainland [the only province with regular wintering?], Batanes, and Zhejiang [significant for migrants? – so how about habitats, trapping there?].


eBird records for 2022 add up to 800 individuals seen

I’ve also emailed Birdlife:

Just copied and pasted the eBird records for [Japanese] Yellow Bunting last year.

Took out several obvious duplicates; and sum is 800 – which to me is best figure for known world population, without guessology.

Yes, should be a lower bound for the population; can figure there must be more. 

[while also, still some duplicates; and can’t be 100% sure that individuals seen in one part of range not also reseen elsewhere]

Top rows of eBird records of [Japanese] Yellow Bunting for 2022 sorted by numbers of birds sighted; just six records in single figures, and vast majority of sightings involved less than 5 individuals

Patchy breeding distribution makes extrapolation challenging

As I prepared to update this, also found a paper published in 2022, chiefly analysing data from surveys conducted from 1998 to 2002, focusing on areas within the known breeding range of this bunting.

This revealed that Yellow Bunting has a very patchy distribution, which surely makes it tougher to extrapolate from the small numbers – occurring in 26 of 30 transects, with a mean density of 0.5 per hectare – found in a single study by Deguchi et al published in 2017 (first online in 2016). From that 2022 paper:

Yellow Buntings were detected in 456 (24.7%) of the 1,846 grids across 22 prefectures, during the study period. 

Our study highlighted the location of the core areas of the Yellow Bunting’s breeding distribution, and demonstrated that lower temperature and deeper snow cover are the critical determinants of those areas. There may be abundant prey and fewer avian competitors, such as resident bird species, in colder regions, and more frequent snow avalanches and water from melting snow may establish the marshy and bushy habitat preferred by Yellow Buntings.

Temperature and snow depth explain the breeding distribution pattern of Yellow Bunting Emberiza sulphurata

Hence it seems this bunting has a very interesting main niche for breeding – perhaps evolved to exploit Japan having steep mountainsides with forest along with extremely heavy snowfalls that result in frequent landslides.

But also, as authors indicate, suggests there may be some vulnerability to climate change.

As for extrapolating population assessments based on studies at a few sites, see a new [updating in June 2023] paper, Pitfalls arising from site selection bias in population monitoring defy simple heuristics. This notes, for instance:

when a population declines, lower-quality sites will be vacated before higher-quality sites. Population declines would be apparent first at the lower-quality sites. Under preferential sampling, however, the best sites are monitored and therefore population declines would only be perceived once the sites of lower quality are vacated. Preferential sampling might therefore delay the detection of a population decline.

Leave a Reply

Your email address will not be published. Required fields are marked *