All years aren’t created equally

17 December, 2017

Note: the following analysis is confined to ‘Continental’ portion of ABA area, and employs the current (8.0.1) ABA checklist. It also borrows heavily from the ABA Rarities Report blogs, eBird data, and the Big Year data sets of birding historian Joe Lill.

In these days of growing interest in ABA Big Years, participants are becoming more strategic in their planning, and more focused on their results. It’s a ‘given’ that the serious contestant is likely to tick all of the 672 ‘common’ species (Codes 1 & 2) from the ABA checklist by the dreaded end-of-year buzzer. The real measure of success is the number of ‘coded’ species (Codes 3-5) that are ticked from that same checklist. Unfortunately for some contestants, not all years will provide equally generous opportunities to rack up impressive lists of rare bird species - irrespective of how much planning and effort is applied. Timing, it seems, can be an overriding factor in determining the outcome of an ABA big year. A comparison of species numbers reported on eBird on a year-to-year basis would seem to be a first step in any attempt to find correlations between 'good' and 'bad' years for twitching, and corresponding climatic trends such as to El Nino and La Nina events. Current predictions are for a La Nina event on the near horizon. And yet to date (so far as I know) there hasn't been any attempt to quantify the annual 'rarity-richness' of years gone by. 

My interest in this matter has recently piqued while following the progress of the leading five 2017 big year birders (Yve, Ruben, Victor, Gaylee and Richard). All have already chocked up impressive year-lists with a bit under two weeks to go. To get a better grip on the relative 'birdiness' of recent years, I decided to start by quantifying the annual eBird reports for coded species from 2011 (John Vanderpoel’s Big Year) to date. How does this year compare to last year? What about in 2013, when Neil Hayward broke Sandy Komito’s long-standing record? To find out, I used the Species Maps section of the eBird Explore Data webpage 
https://ebird.org/ebird/map/ to extract all ABA coded species reports (as per ABA Checklist v.7) on a year-to-year basis. I also consulted with the weekly ABA Rare Bird Report blogs, accessed from the ABA website. I recognize that going back only seven years is not sufficient to draw many conclusions – but maybe someone with more keyboard endurance can take over and go back another five or ten years, perhaps using additional reporting media (including NARBA). For the purposes of this exercise, Hawaiian sightings have not been considered, as that would add more confusion than clarity to the sorts of trends we’re trying to identify.

In a perfect world, the number of coded species reported on eBird in a given year could be regarded as a reflection what’s going on in the field – e.g. if 2016 had 37% more rare species reported than did 2012 (and it did), then it would be nice if we could assume that there were about 37% more rare bird species bouncing around the ABA area in the later year. But it isn’t that simple. Other factors probably include the ever-improving bird-reporting systems – with increasing use of eBird, the increasing interest in birding (resulting in more birders stomping around out there - hopefully occasionally encountering unusual birds), increased strategically-timed pelagic birding opportunities, and possibly improved coverage of the Alaskan outposts by dedicated birders and birding tours (though it seems that numbers of birders venturing to the islands substantially swell and wane on an annual basis in reaction to the relative successes of the previous year). Taxonomic changes present an additional consideration - at least in the longer term, with the net effect of all those splits and lumps gradually swelling the number of prospective coded bird possibilities.

After crunching the eBird numbers, I formulated two different representations of the resulting summaries. Table 1 provides a straight-up year-by-year comparison of the number of rare bird species reported at least once on the eBird registry and/or the ABA Rare Bird Report. Further below, in Table 2, I’ve attempted to differentiate between reported rarities that the serious big year birder might have had a good chance of sighting at some time during the year, and those species that couldn’t readily have been converted to ‘ticks’ by ambitious twitchers on any occasion that year (e.g. one-day-wonders).

Table 1. Summary of the number of ‘coded’ bird species reported annually on eBird from 2011-2017*

2011
2012
2013
2014
2015
2016
*2017
Code 3
74
68
75
71
73
79
78
Code 4
49
34
43
39
47
53
53
Code 5
15
9
14
11
19
19
25
Total:  
136
111
132
121
139
151
156








* 2017 figures are current only to time of writing – 17 December. Will update at end of year.

As suggested earlier, I thought it might be interesting, at least to my fellow big year birders, past, present, and future, to categorize the rarity reports from each of the last six years on the subjective basis of ‘twitchability’. Obviously, not all reported rare birds can be chased down by even the most cashed-up and reactive birders; many are observed only fleetingly - perhaps with a single sighting by a single birder. Hence, the total number of rare species reported in a given year is not the same as the number of rare species accessible to that year's listers. It doesn’t really help if a Code 5 Eurasian Wryneck shows up on a no-access military installation off the California coast, or if an Atlantic Yellow-nosed Albatross is seen only on a single pelagic trip out of New York. At the other extreme of the ‘twitchability’ scale, a Tufted Flycatcher in Madera Canyon with a months-long history of almost daily sightings in a heavily-birded public place, is pretty much a ‘dead cert’, as the Australian birders would say. 

For the sake of satisfying my own curiosity, I created a practical set of criteria to try to distinguish between incidences of ‘twitchable’ versus ‘un-twitchable’ individual rare birds, by screening thousands of eBird reports generated from 2012-17. An inadequate description and justification of those criteria is provided further down the page. 

Table 2. Number of rare bird species reported in each of the past six years are separated into two groups – ‘twitchable’ and ‘untwitchable’. I ran out of steam before breaking down the 2011 reports!

2012
2013
2014
2015
2016
2017*
untwitchable
29
36
35
29
35
36
twitchable
83
96
86
110
116
120
total
111
132
121
139
151
156

To qualify for ‘twitchable’ status during a given year, a species must have presented the keen twitcher with at least one opportunity for a fair shot at encountering it. In most cases, this was established when, after a report was made, subsequent day-reports were timed in a manner conducive to a successful twitch. My subjective definition of 'conducive timing' is that when a cluster of two or more day-reports for an individual bird occurs, the bird was deemed twitchable any time when two day-reports occurred with no more than two 'reportless' days between them. This may still call for a madman or rabid-woman to take a chance of success, but its at least possible. The idea with the prescribed maximum period for no sightings is that that after two days of dipping, the twitcher could be forgiven for moving on. In those instances where sightings were confined to two back-to-back days, I erred on the ‘untwitchable’ side.

Of course, in practice, many birders ultimately rely upon their gut feelings about the ‘stickability’ of freshly-reported birds to decide whether or not to hightail it to the airport. Some, including myself during non-'big'-years, will automatically invoke the ‘two-day’ or ‘three-day’ rule (that's me) – wherein the bird needs to be relocated for a second, or perhaps third day before flights are booked. Predominant factors that affect the decision to jump - or not, include costs and inconvenience. For big year birders however, the element of desperation can understandably become an overriding factor. The especially motivated player can expend big efforts and blow tight budgets adopting the ‘fly first and ask questions later’ approach, only to accumulate a soul-destroying string of dips (it's true, believe me). 

For the purpose of this exercise I’ve approached the twitchability issue for pelagic bird species a bit differently. Coded tube-nosed species that typically show up at least four or five times per year off the California coast or the Atlantic Gulf Stream during the peak seasons (Spring in the east, Autumn in the west) are regarded as twitchable - even if it might require up to a half-dozen day trips to secure a sighting. Examples of these include Flesh-footed Shearwater, Buller's Shearwater, Least Storm-petrel, and Craveri’s Murrelet on the west coast (it required 14 day trips and a 8 wishful whale-watch missions [which never gained me a species!] for me to secure all four of these); and Great Skua, Fea’s Petrel, and Trindade Petrel (never did get this one) on the east coast.

Due to the limitations of this little study, there aren’t any earth-shaking conclusions to draw beyond the fact that there does appear to be significant year-to-year variation in the number of rare bird species reported in the ABA area. The annual rates of reported rare species and ‘twitchable’ species share similar curves, suggesting that the simpler, less subjective indicator of ‘good’ vs ‘not as good’ years for big year birders might be a straightforward index of reported coded species per year. In other words, potential for epic Big Year results can in part be measured, though sadly in retrospect, by the number of coded species reported on eBird that year. And so, with the benefit of 20:20 hindsight, it looks like the current year of 2017 has eclipsed the previous year as the best twitching year to date. That's gotta be the peak, right? But then, a 'half-glass-full' interpretation of the data could be that the upward trajectory of rare bird species numbers in the ABA area over the past four years indicates that 2018 is likely to be even bigger!  I'll let you know how full that glass really was in about twelve months time.


Maybe it was all about El Nino after all...
Looking at Diagram 1 taken from the NOAA weather data (below), it looks as if the bumper years of 1998 (Sandy Komito’s massive Big Year), 2011, 2016 and 2017 for bird rarities in the ABA area all fell at least partially in La Nina conditions, either immediately behind El Nino events (2011, 2017) or straddling the transition from El Nino to La Nina (1998, 2016). But unless we can at least partially dismiss the 2013 and 2015 milder elevations in rare species numbers as more a reflection of an ongoing increase in reporting efforts, than an increase in rarity diversity, the correlation doesn't really work...









Bonus table: It was an easy exercise to compare the rarity counts of more recent Big Year birders as a proportion of the rare birds reported on eBird at least once during ‘their’ year, again borrowing from the mine of data curated by Joe Lill. At some stage I will update this comparison to include earlier years, and final results for 2017.

Year
Birder
Rarities in
Rarities by
Percentage
ABA area
Birder
2011
John Vanderpoel
138
83
60%  
2013
Neil Hayward
132
83
63%
2016
John Weigel
151
111
74%
2016
Olaf Danielson
151
107
71%
2016
Laura Keene
151
94
62%
2016
Christian H.
151
82
54%
2017*
Yve Morrell
156
87
56%

*As at 17 December, 2017, on the homeward stretch, Yve is a nose ahead of the closest of the four other Big Year birders (Ruben Stoll, Victor Stoll, Gaylee Dean and Richard Dean).