Bird Lists: A rather freaky THEORY…
***WARNING - sums ahead***
Funny things bird lists. We all probably keep them…life lists, year lists, county lists, patch lists. Some of us eat up inordinate amount of miles, time and money tearing about the country in search of the latest mega, whilst others of us are content to stay pretty much on our local patch unless there’s something close-ish to tick and there’s not much else going on. Birders are very mixed bunch, with very mixed resources and very mixed expertise and so you might think that there would be no way to predict how many species any given one would see over any given time and at any given place - BUT how about if you could? How about, for example, if you could figure out how many new patch ticks you’d get next year or how long this years’ list will end up or even how many species you’ll add to your current list between now and Christmas? Well, I reckon it might just be possible…
FIRST a quick recap on the Law of Diminishing Returns (see Star Trek, Pokemon and Relativity posting for the full thing).
“Imagine - it’s January 1st and the new, year list beckons. You rack up a respectable number of species and go to bed happy with the start to the year. As luck has it, you get out the next day too. You quickly find though that you’re seeing many of the same species and relatively few new ones. The following weekend arrives and you go out again…same story. In fact as the year progresses (apart from slight peaks in the spring and autumn) last year’s easily beatable total begins to seem less and less assailable because whilst your first 10 species took 5 minutes to bag, your last 10 look likely to take 5 weeks or 5 months! The weird thing is that even if you widen your search you’ll still hit this same wall… and what’s more, the same holds true for bird races, county lists, local sites, life lists, foreign trips and world lists. The first birds are easy to get, the next much harder, the next harder still - the only thing that changes is the scale”.
If you plot a graph of this phenomenon, i.e. number of NEW SPECIES seen (vertical axis) against TIME interval (could be hours, days, weeks, months, years) on the horizontal axis, you’d get something like this…
You see loads of new species early on, proportionately fewer as time goes on.
Now 'interesting' as this graph is, there is a better one because you can plot the same info in a different way and show the ACCUMULATION of species over time – much more useful when you’re trying to add species to your list! Imagine it’s the end of the year and you managed to see 200 species. Plotting such a graph would let you see how quickly you accumulated them over the year – a kind of running total if you like, provided of course that you kept a note of when you added each new species to your list. For example, you may have seen 96 species in Jan, another 24 in Feb (makes 120), another 21 in Mar (makes 141) and so on. Now sad git that I am, I did this for my 2003-2006 north-west year lists. I saw a few more birds some years than others, so I converted the actual species numbers to percent of total species seen in each year so that they were on the same scale, so to speak. The numbers on the bottom axis are just the year chopped up into 2 month segments (1 = the first seconds of the New Year when I had no birds and then 2 = Jan/Feb; 3 = Mar/Apr; 4 = May/Jun etc.). I was VERY surprised at what I found (see below).
some kind of cosmic pattern to my birding. EVERY YEAR I was seeing 60% of my birds in Jan/Feb and had ticked 85% of my year tally by the end of April. Thereafter, species were added more and more slowly (UNsurprising), but in a predictable way (very surprising). What this means is, it is possible to predict the final length of your year list based on the % of birds you see early in the year (!), assuming your patch effort each month (e.g. no.visits) stays the same, give or take a few.
This is pretty cool, coz it means you MIGHT in theory be able to read off where you are in the year from such a graph and see what percentage of your PATCH finds are in the bag, and how many more you might expect to get. That’s assuming it’s some kind of mystical Universal Law of Bird Lists whereby you can extrapolate from the whole of the north-west down to your local patch. Question is, does it work? Let’s see what happens if you apply the theory to, in this example MY local patch. I've tidied up the above graph to make it a bit easier to follow (see below)...
Here goes for 2010 Moore ticks then! The dashed line on the graph shows where the end of March is in the year (i.e. 25% of the year gone), whilst the solid line shows where the end of June is (i.e. 50% of the year gone). According to this, at the end of March I should already have seen about 70% of what I’m gonna see this year on the patch and by the end of June about 90%! OK, I’m on 125 species at the moment and it’s almost end of June, so if 125 species is 90% of what I’m GOING to see, my final year list should be 125/0.9 = 139. Cool! Right, if that’s the final list length then (100% of species I'm going to see) what SHOULD my list have been on at the end of March? This is a check that the theory works lol. By the end of March the theory predicts I should have seen 70% of this year’s patch birds, or, 0.7 x 139 = 97. Checking back, I was actually on 96 – pretty damn close! I reckon I’m on to something here… happy days there are still 14 species up for grabs on the patch! Well, not quite, because I know from previous years I always miss about 7% of what’s around, so that’s 0.07 x 139 = 10 birds, that have to come off realistically speaking, which leaves my likely final predicted year tally at 139-10 =129. DOH! That means I'm only likely to find just FOUR more species by the end of the year – meh. Reality sucks eh? Mind you - Bittern and Golden Plover will be two of them... wonder what the OTHER two are gonna be :)
OK, so if you’ve stayed awake thus far you should be quite intrigued by now because (a) the rate at which birds accumulate on year lists when birding the entire north-west seem to be similar between years, (b) it seems to be predictable as well and (c) seems to apply equally well to local patches! This is like fractals… How far can we drill down I wonder? Happily, I have LOADS of data from Moore - even SADDER git… yes, accepted - BUT it means I can see if the theory works on an even SMALLER scale! Let’s go for 6 months worth of obs at... Pumphouse Pool (a little bit of Moore)... and look just at... WeBS species visible from the hide. What happens when you plot THAT tiny subset on the same graph as the north-west stuff? Does the same predictable pattern appear???
"Cor blimey, says I. Now ain’t that a cosmic coincidence? Be it just I..." (*remains in Pirate mode*) "...or be those two curves pretty well chartin’ the same course?"
What IF I go even smaller and sit in one place and record the accumulation of new species over a couple of hours? Will the same pattern hold true?? Now, I can’t physically do this just atm, BUT I can look at Misty May Morning on here… that was a sit and watch session AND I noted down the times… should be able to test the theory from that!
Right... *rubs hands in anticipation*
Started at 06:20 and finished at 08:45, which is… 145 minutes, or about two and a half hours. Let’s see what birds I got then. Need to chop the whole sesh up into time chunks and see how many species were seen in each chunk. Easy. Let’s go for five chunks of 29 minutes each, or 20% of the total time, 40%, 60%, 80% and 100% - makes the sums nice an easy...
06:20 - 06:49 Shelduck, Grey Heron, Mallard, Magpie, Canada, Tufty, Reed Warbler, Gropper, Willow Warbler, Wren, Mute Swan, GC.Grebe, Sedge Warbler, Mistle Thrush, Blackbird, Cormorant, Lapwing, Carrion Crow, Greenfinch, Song Thrush, Pheasant, Stock Dove (22 species).
06:49 – 07:28 Blackcap, Great Tit, Coot, Black-headed Gull, GBBG, Oystercatcher, Reed Bunting, H.Gull, Long-tailed Tit, Whiethroat, Chaffinch (11 species)
07:28 – 07:57 Goldfinch, Bullfinch, Swallow (3 species)
07:57 – 08:26 Dunlin (1 species)
08:26 – 08:45 Sanderling, Sand Martin (2 species)
That’s a grand total of 39 species. SO, I saw…
22/39 = 56% of species by the end of the 1st 29 minutes (20% of the time I was there)
33(22+11)/39 = 85% of species by the end of the 2nd (40% of the time)
36(33+3)/39 = 92% of species by the end of the 3rd (60% of the time)
37(36+1)/39 = 95% of species by the end of the 4th (80% of the time)
39(37+2)/39 = 100% by the end
Here’s how those five data points compares with the theoretical graph!
Well, I reckon that’s pretty fuckin amazing tbh… it may ACTUALLY be a Universial Law of Birding Lists, he, he \m/
Now all I need to know is whether the theory works on YOUR lists... IT SHOULD!!!