Identification of Olive-backed and Tree Pipit by call

Introduction

Every autumn, there are several records of Olive-backed Pipits in Europe. Up until 1987 there were just 61 records throughout Europe. Today Great Britain has more than 300 records on its own (almost 50 records in 2012) which even led the British Birds Rarities Comitee to remove it from the BBRC list.

Surely one reason for the increase of the records is better observer knowledge due to improved field guides. But it is not known if there has been an expansion of the breeding grounds to the west. It is also unknown if there is a small wintering population in southern Europe or northern Africa. Records of wintering birds or in spring are extremely rare.

Almost all records are of seen birds – and seeing them isn’t easy: Olive-backed Pipits often hide in the vegetation. This explains why most records are from small islands with little vegetation and many observers, for example Heligoland (about 75% of the German records) or the Shetlands (about 50% of the British records).

The number of records from the rest of Europe would certainly increase if it was possible to identify Olive-backed Pipits just by call. In fact, this is a bit complicated: No Bird Rarity Comittee will accept a record of a (just heard, not seen) bird without a sound recording – but it is actually not really established how to recognize a Olive-backed Pipit from a sound recording. I encountered the same problem when I recorded a pipit on the German Isle of Sylt in October 2008. It sounded like an Olive-backed, the sonogram looked like an Olive-backed – but how to be sure? Finally Magnus Robb confirmed the record and the bird was accepted by the German Rarity Comitee (DSK).

Since 2008 there are some hints about its identification on the internet and in the literature. Some of them are wrong, some could be confirmed in my analysis. With this blogpost I will try to provide everybody with the necessary information to identify a possible Olive-backed Pipit.

Geografical  variation:

The Olive-backed Pipit Anthus hodgsoni has two subspecies: A. h. yunnanensis breeds in the northern part of its range as far west as the Ural. The other subspecies A. h. hodgsoni breeds in the Himalaya, in central China up to the 38°N and in Japan. It is unknown which subspecies breeds in Korea (Alström 2003). Vagrants to Europe belong to ssp. yunnanensis. Both winter in India and South-east Asia. The wintering grounds overlap completely and there are only few sound recordings from the breeding grounds. So I didn’t differentiate the recordings between subspecies.

Waldpieper

Olive-backed Pipit, 21.10.2012, Heligoland, Germany. Olive-backed Pipits often hide in the vegetation and are difficult to watch.

Tree Pipit Anthus trivialis also has two subspecies: A. t. trivialis breeds throughout Europe (apart from the very south) and winters in Africa and India. In the east it is replaced by A. t. haringtoni. Some authors state that this subspecies already occurs in Turkestan/Kazakhstan, others say it breeds eastwards from Kashmir; these birds winter in India. Shirihai (1996) additionally recognizes A. t. schlueteri. It breeds in the mountains of western central Asia and winters in India, the Middle East and in Eastern Africa. However Alström et al. 2003 don’t recognize this subspecies as valid. In summary, the Tree Pipits of India include at least one other subspecies. Therefore I analyzed them separately.

Baumpieper_Adult

Tree Pipit Anthus trivialis trivialis, 30.05.2013, Lkr. Wörth, Germany.

Data set

For this article I analysed 190 calls of 56 recordings of Tree Pipits from the western part of the range, 17 calls of 4 recordings of Tree Pipits from the eastern part of the range and 190 calls of 48 recordings from the whole range of Olive-backed Pipit.

Call types

Both species have several different calls. The alarm call is easily distinguishable: it is a „sit“ call, that is repeated in short intervals. In Olive-backed Pipit it is a horizontal stripe in the sonogram and is clearly higher pitched than Tree Pipit (about 5-6 kHz for Tree Pipit and about 7 kHz for Olive-backed Pipit). Alarm calls of Tree Pipits look like a hook in the sonogram and sound different. You will hardly hear any alarm calls of Olive-backed Pipits in Europe.

The typical flight calls are more difficult to distinguish from Tree Pipit. If you have a good recording, identification should always be possible with this article. There are some similar calls (sonogram 1, recording 2-4 and sonogram 2, recording 3-5) that are even more difficult to distinguish.

Rarely the birds use in combination with the typical flight calls another call, very similar to the alarm call (or even the same. See sonogram 1 & 2, in each case the 7th recording), but the call is not combined to series as the alarm call. Tree Pipit once again is much lower pitched and v-shaped in the sonogram, Olive-backed Pipit is higher pitched and in the sonogram constantly falling or only with a small hook. It pays off to listen for this call in migrating birds as it is a very distinct call.

Baumpieper_Rufe

Sonogram 1: Different calls of Tree Pipit: 1. Recording: Typical flight call, 21.04.2011, Kirschgartshausen, Germany © M. Feuersenger; 2. Recording: flight call, more rarely used than the first type, 28.08.2012, Landsort, Sweden © A. Dalton; 3. Recording: Flight calls with such a slow modulation are rare and there seems to be no transition between the fast and the slow modulated calls. 12.04.2013, Jablonna, Poland © J. Matusiak; 4. Recording: Rarely, the flight calls even rise in pitch (1. call in the recording). 21.04.2011, Kirschgartshausen,Germany © M. Feuersenger; 5. Recording: Alarm calls, 25.06.2013, Yvelines, France © J. Rochefort; 6. Recording: Alarm calls, 4.08.2013, Rimmelsberg, Germany © V. Arnold; 7. Recording: There is a rarely used, completey different flight call which is similar to the alarm calls. 15.09.2012, Kwintelooijen, Netherlands © P.G. Gelderblom

Waldpieper_RufeDifferent calls of Olive-backed Pipit: 1. Recording: Typical flight call , 05.01.2013, Da Lat area, Vietnam © P. Aberg; 2. Recording: Typical flight calls, 11.04.2013, Tangshan, Hebei, China © M. Slaymaker; 3. Recording: These calls are analog to the calls of Tree Pipit (sonogram 1, 2. Recording), but Olive-backed Pipit doesn’t use them as often as Tree Pipit. 26.09.2010, Happy Island, Hebai, China © D. Marquez; 4. & 5. Recording: These flight calls are analog to the calls of Tree Pipit (sonogram 1, recording 3 & 4). They are often distinctly higher than the others and easy to identify by pitch. 6.12.2011, Doi AngChang, Thailand © T. Linjama; 6. Recording: Alarm calls, 28.07.2010, Onon Gol Camp, Mongolia © C. Bock; 8. Recording: Just like Tree Pipit, Olive-backed Pipit has another, rarely used flight call. It is very distinctive. 03.01.2010, Uttar Pradesh, India © P. C. Rasmussen;

ID pitfalls

Very bad recordings could be confused with Redwing Turdus iliacus in the sonogram, but shouldn’t pose real id problems. Red-throated Pipit Anthus cervinus has the same pitch as Olive-backed Pipit. The call is hardly modulated and therefore shows as a quite constant thick line in sonagrams. Calls of Red-throated Pipits sound much more squeezed than those of Tree and Olive-backed Pipit.

Rotkehlpieper

Sonogram 3: Red-throated Pipit, flight calls, 01.03.2013, Larnaca Sewage Works, Cyprus © J. Honold

Differences in the flight calls of Tree and Olive-backed Pipit

There seems to be some wrong information about the calls of Olive-backed Pipits. The most important is: Tree Pipit never exceeds 8 kHz. But there are definitely some out there (for example sonogram 7, 2nd recording). To identify a call just by the highest frequency is very difficult because the amplitude of the call (and thereby maximum frequency; note that amplitude here does not refer to loudness but to frequency range of the call at one point in time) varies with the quality of the recording. It also varies with the position of the sharpness controller in Raven. It just helps for a fast classification of the recording: only few flight calls of Tree Pipit are higher pitched than 8.5 kHz with the standard sharpness setting of 250 in Raven.

Due to the big differences in the quality of the recordings, I looked for suitable methods to compare them. As reference points, I chose the mean starting frequency (MSF → Point 2 in sonogram 4) and the gradient of the call (blue line in sonogram 4). There is an animation how to measure MSF and the gradient HERE. Caution: the gradient was measured from back to front to prevent an algebraic sign chaos. So a positive gradient means, the call is falling, a negative gradient means the call is rising. With the software “Raven” you can easily measure these two values. If you position your mouse at the points, you can see the frequency and point in time at the bottom of the window.

Olive-backed Pipit_measure_pointsSonogram 4: Measuring points in the flight calls of Tree- and Olive-backed Pipit. The blue line shows the gradient of the call and connects point 2 (mean starting frequency) and point 3 (mean ending frequency). The blue line follows the center of the modulations. Point 2 is positioned at the first point in time where it is in the centre of the modulations (yellow marked. Left of the yellow marking, the modulation isn’t symmetric). It is not relevant at which point of time point 3 is placed because the gradient remains the same. It is recommended to place it in the right part of the call to increase the accurateness of the result. There is not always a point 4 (especially in Olive-backed Pipits). For an animation how to measure, click HERE.

Another good character is the duration of one modulation. Tree Pipits modulate considerably faster. This is the reason why single modulations in worse recordings of Tree Pipit often cannot be resolved. Usually Olive-backed Pipits have about 10-15 modulations per call, Tree Pipits about 15-25. But the call length varies a lot and therefore is not characteristic.modulation countingSonogram 5: Instruction on measuring the duration of a modulation at the example of a Tree Pipit. In the left sonogram, the sharpness setting in Raven is at 250 (standard). If you use a lower setting (170 in the right sonogram), thereby increasing time resolution, you can often recognize and count the modulations better. Beware of calls as in sonogram 1, recording 3 & 4 or sonogram 2, recording 4 & 5! They are of a different call type and much slower modulated. They have durations of more than 14 ms for one modulation and are easily recognizable. For an analysis, you should recognize at least five distinct modulations. Measure the time from tip to tip of the modulations. If you count the modulations between these two tips, you have to subtract one modulation, because you measure just the half of the first and the half ot the last modulation (in the sonogram above 22 modulations have a blue dot, but you measure just the time of 21 modulations). Very rarely there are two clearly different modulation frequencies in one call. Measure the faster modulation frequency in these cases.

BaumWaldpieper

Sonogram 6: Two sonograms of Tree Pipit (left) and Olive-backed Pipit (right) with a larger time scale (compared to the other sonograms) to show the modulations. Olive-backed Pipit modulates clearly slower. Both lower sonograms show extreme birds where the species overlap, the upper ones are typical. 1. Recording upper left: 01.05.2009, Hammarsudden, Sweden © P. Åberg; 2. Recording lower left: 28.09.2012, Heligoland, Germany © J. Honold; 3. Recording upper right: Very small gradient for an Olive-backed Pipit but everything else is typical. 26.09.2010, Happy Island, Hebai, China © D. Marquez; 4. Recording lower right:   Japan © Y. Endo

When measuring these values, the following problems can occur:

  • some calls (especially of Tree Pipit) have a curved decrease to point 4 (example sonogram 7, 4th recording). To measure such a recording, you have to set the MEF further left. Stay away from the curved part and set point 3 at about two thirds of the call.
  • in calls as in sonogram 1, recording 2 and in sonogram 2, recording 3 it is not possible to measure a gradient. If the duration for one modulation and the MSF is not typical, don’t try to identify the birds. Tree Pipits don’t use these calls often and Olive-backed Pipit even less.
  • very short calls. You can just measure MSF (rarely you can measure the gradient). If it is not characteristic, you shouldn’t use them for identification.

→ if the MES of these different call types is not characteristic, you should concentrate on the typical flight calls (like the first recording in sonogram 1 and the first and second recording in sonogram 2).

For an identification, it is helpful to combine these values. So I defined parameter 1, which is the product of MSF and the gradient. You can measure these two values even in the worst recordings. But using just these two values, there is still some overlap. I defined Parameter 2 as the product of MSF, the gradient and the duration for one modulation.

To help the reader, Johannes created a worksheet. Just type in your measured values of your recording and Excel calculates the parameters for you. If you are interested in all the formula, you can find them in the appendix.

Worksheet for the calculation of the parameters. There is already an example pipit entered which you can overwrite. For the time durations, type in the time points and the worksheet will give you the duration. You might as well type in the duration if your software allows you to measure it directly. If you can’t measure some values type 0 into the worksheet! If you want to get the arithmetic average of three calls, scroll down and you will find two more worksheets. Below them, you will find a table that gives you the average values of these three calls.

You can compare the values with table 1:Table_1Table 1: Results of this analysis. Every data set stands for one call. Caution: the gradient was measured from back to front to prevent an algebraic sign chaos. So a positive gradient means, the pitch of the call is falling, a negative gradient means the call is rising. If some measurements are out of the given range, you shouldn’t identify the bird – even if other measurements seem to be diagnostic!

The calls within one recording vary a lot and so there are some outliers in table 1. Even with parameter 2 there is an overlap. The results improve if you measure more calls of the same bird (at least three) and calculate the arithmetic mean. You can compare the results with table 2:Table_2Table 2: Instead of using single calls for a data set, in this table every data set represents the arithmetic mean of the calls of a recording. With this method there is no overlap at all. Caution: the gradient was measured from back to front to prevent an algebraic sign chaos. So a positive gradient means the call is falling, a negative gradient means the call is rising.

In the tables you can see that the decrease in pitch in Olive-backed Pipit calls is faster (= a greater value for the gradient), the pitch is generally higher and the call is slower modulated. Some Tree Pipits have a very high pitch, but can be identified by the other characters.

Interestingly, the Tree Pipits of the eastern part of the range show a higher MSF and a higher value of the gradient in table 1. So the values are more similar to Olive-backed Pipit but the sample size is very small. This is the reason why this population is not separated any more in table 2.

More differences

There are some more differences, but they seem to be very variable and not diagnostic:

  • the pitch of Tree Pipits falls slower. So sometimes the highest frequency is not at the beginning but in the middle of the call. At times the pitch even increases.
  • Tree Pipits often show point 4 (Olive-backed Pipit rarely do). They abruptly fall down at point 3 to point 4. Sometimes, they change their gradient slowly to the end of the call. The call gets a swinging shape like a music note (see sonogram 7, 4th recording).
  • on average the amplitude of Tree Pipits is bigger than in Olive-backed Pipit. (again, speaking about frequency range in one point in time)
  • Tree Pipits almost always show the same amplitude (see above) in the main part of the call. So the sonogram shows a broad, falling stripe. In Olive-backed Pipits recording’s, the amplitude sometimes decreases towards the end of the call (so it is less rough). But the variation is tremendous – so it is not possible to identify a bird by this character. There are even Olive-backed Pipits in which the amplitude increases (sonogram 8, 2nd recording).
  • in Olive-backed Pipits point 1 seems to be – also relatively – higher than in Tree Pipits. But the point changes with the recording quality. It could also be caused by a higher willingness of the sound recorders to upload recordings of Olive-backed Pipits with a worse quality than recordings of Tree Pipits.

Baumpieper_Rufe_Variation

Sonogram 7: Variation of the typical flight calls of Tree Pipits. Look at the differences in pitch, the small gradient up to the 3rd point, the fast modulation and in all recordings the call falls at point 3 down to point 4. In the 4th recording you need to set MEF at the half or 2/3 of the call. 1. Recording: 01.05.2009, Hammarsudden, Sweden © P. Åberg; 2. Recording: 21.03.2013, Cape Pyla, Cyprus © J. Honold; 3. Recording: 21.04.2011, Mannheim, Germany © M. Feuersenger; 4. Recording: 02.01.2010, Chambal Safari Lodge, India © P.C. Rasmussen; 5. Recording:  This call is strange with the bend in the middle. There would be no problem in identification, but it seems to be more safe to exclude such calls from an identification.16.09.2012, Zvenigovsky District, Russia © A. Lastukhin.

Waldpieper_Rufe_VariationSonogram 8: Variation of the typical flight call of Olive-backed Pipits. Only in the 5th recording the frequency falls from point 3 down to point 4 (in the other recordings, there is no point 4). The modulation  is very slow. In the 4th and 5th recording the amplitude gets smaller towards the end of the call, sometimes it stays the same and in the 2nd recording even increases. All of these recordings have a quite small amplitude. There are also Olive-backed Pipits with a greater one out there. 1. Recording: 11.04.2013, Tangshan, Hebei, China © M. Slaymaker; 2. Recording: Egana, Israel © Y. Perlman; 3. Recording: The amplitude of the calls get sometimes smaller to the end. 03.01.2002, Patna Bird Sanctuary, India © P.C. Rasmussen, Source4. Recording: 02.11.2012, Linosa, Italy © M. Vigano; 5. Recording: In a few cases, Olive-backed Pipits show also a drop down of the frequency at point 3. But it doesn’t get much more distinct than this. 01.01.2002, Chambal Safari Lodge, India © P.C. Rasmussen Source.

How do you get aware of a potential Olive-backed Pipit?

With much experience or in direct comparison you can hear that Olive-backed Pipit calls are higher pitched. Sometimes you can hear the conspicuously fast decrease in pitch. Tree Pipit also sound rougher (thanks to the on average greater amplitude) and sometimes Olive-backed Pipit looses its roughness towards the end of the call.

It pays off to listen for the inconspicuous, bunting like „sit“ calls of migrating birds as they are very distinctive.

Calls of Olive-backed Pipits bear more resemblance to Red-throated Pipit. This is easily explained: Both have the same pitch, the same gradient and Olive-backed Pipit doesn’t sound as rough as Tree Pipit.

There are also great differences in the phenology between the two species: most Tree Pipits migrate in September, late individuals in the begin of October. In the middle of October Tree Pipits become rare. The very first Olive-backed Pipits migrate in middle of September but most birds reach Europe in October. So it is highly recommended to be ready to record every Tree/Olive-backed Pipit from mid October onwards.

Discussion 

For this analysis I used recordings from http://www.tarsiger.com, http://www.xeno-canto.org, avocet.zoology.msu.edu, http://www.youtube.com and macaulaylibrary.org. The problem is that you don’t know the sound recordists personally and can’t assess the level of their knowledge. The “problem calls” of the two species are flight calls and it is difficult to identify a flying bird with the binoculars AND take good recordings of it.

However, the recordings could be clearly divided into two groups. Just three recordings didn’t fit into these groups. Two of them belonged to calls of Red-throated Pipit (I asked the recordists and they agreed that they were Red-throated Pipit) and one recording of India titled as Olive-backed Pipit showed without any doubts a Tree Pipit. But to be safe, I didn’t take this Pipit into the analysis.

A little imprecision is given by the fact, that different people will measure in a different way. So feel free to post your results or send them to me.

Summary

With the help of this article you are able to identify flight calls of Tree Pipit and Olive-backed Pipit by pitch, gradient and duration of one modulation. If your recording isn’t good enough to measure all these values, it is still possible to identify most birds.

I am aware that there is quite a lot of maths in this article. But this is the only possibility to identify these two variable species. I hope everything is understandable and nobody will be put off by this.

Thanks to all who uploaded their recordings in the internet or sent them to me!! Without you this analysis wouldn’t have been possible!

Special thanks go to Johannes Honold and Peter Schleef for interesting discussions about the topic, their help and hints to the manuscript and for sending recordings of and discussing about some “problem birds”. And thanks to Magnus Robb for identifying my first Olive-backed Pipit and also for his informative answers.

Appendix:

Formula 1:

GradientFormula 2:

Parameter_1Formula 3:

Time for one modulation

Formula 4:

Parameter_2

Literature

Alström, Mild & Zetterström (2003): Pipits and Wagtails of Europe, Asia & North America. Christopher Helm Verlag

Magnus Robb in Deutsche Avifaunistische Kommission (2012): Seltene Vogelarten in Deutschland 2010. Seltene Vögel in Deutschland 2010: 10-49

Helbig, A. J. (1987). Feldbestimmung des Waldpiepers Anthus hodgsoni und sein Auftreten in Europa. Limicola 1, 73-85

Shirihai (1996): The Birds of Israel. Academic Press Limited, London

Addendum: Corrected several grammatical mistakes but no change of content. 20.11.2014

2 Antworten zu “Identification of Olive-backed and Tree Pipit by call

  1. Pingback: Recente CDNA-besluiten | Avifauna van Gelderland·

  2. Pingback: Taigakirvisen kutsuäänestä | Sounds from Nature·

Schreibe einen Kommentar

Trage deine Daten unten ein oder klicke ein Icon um dich einzuloggen:

WordPress.com-Logo

Du kommentierst mit Deinem WordPress.com-Konto. Abmelden / Ändern )

Twitter-Bild

Du kommentierst mit Deinem Twitter-Konto. Abmelden / Ändern )

Facebook-Foto

Du kommentierst mit Deinem Facebook-Konto. Abmelden / Ändern )

Google+ Foto

Du kommentierst mit Deinem Google+-Konto. Abmelden / Ändern )

Verbinde mit %s