|
Steve,
Thanks for that. Good stuff.
I was going to ask if you could source a
graph pretty much as you have done but you beat me to it. The general
annual pattern is supportive of what I wrote before. It seems to suggest
a possible change in status but
it may be hard to get any significant analysis from
that.
There are some issues to be aware of
though before trying to take the graph beyond its usefulness. This graph is of
general records which includes ad hoc ones, I suspect excluding GBS data, and
not based on a set survey method. I mainly used GBS data. I believe the
real pattern in the suburbs (sampled by GBS) should in reality be very similar
to the real pattern in the wider COG area for this species.
Firstly, that graph shows records for the
first 5 years are only a tiny fraction of those for the other groups of 5 years
since then. However the GBS data (you only need to look at the front cover of
The GBS Report), shows that abundance for the first 5 years is very much
higher, than for the years since then. Clearly what has happened is that the
species probably was much more common in those early years but in those
early years, fewer general records went into the COG system in a way that is
still preserved in our current records and thus this graph. The numbers
1986-1990 are surely high because those years we were doing the COG Atlas.
Indeed I think the proof is there, that the graph for those 5 years looks very
much like the graph printed in the Atlas (3 years included). Other than that
there are far more general records in recent years because our technology has
improved.
Secondly, is the whole aspect of ad hoc
records of a species that is not so rare that all observations go into the
formal record. People have a range of biases and they are to: preferentially
submit a record because it is at an unusual time of year, these tend to
obscure the true trends by highlighting the unusual, or: submit a
record because it is at the time of year that the interesting behaviours,
such as migration, happen, these tend to artificially exaggerate the true
trends, or records have a spike because of special effort surveys such as the
Blitz weekend. There could be a lot of that in the big October spike in the last
5 years, coinciding with the Blitz activity (maybe Barbara could comment).
Analysing the importance of these biases is so hard it rarely is attempted. All
of these make interpreting annual trends difficult.
Thirdly, the demonstration of the annual
pattern by this species from my writing and speaking about it, of itself has
been a feeder into this process of creating sources of bias into what
observations, people decide to put onto the formal record or not.
Lastly, of historical curiosity is that I
created the histogram in my 1991 article that first showed this trend in the
species, from the tiny graphs in the COG ABR in CBN. This was as far as I
know the first use of GBS data in formal national literature outside of COG. In
the 1980s and early 1990s monthly abundance histograms for most species recorded
in reasonable numbers on the GBS were printed on separate sheets and
individually cut out and glued onto the printing proofs, on gaps cleverly
created by setting a margin to fit them. (McComas Taylor was the clever person
who did most of that work. I know because I did a lot of the editing, cutting
and glueing.) I used these tiny printed graphs and measured the height of each
monthly bar for those 6 years then available and then used Lotus 1-2-3 to create
a composite graph, also allowing for that the size of the printed graphs
differed between years. Back then the GBS summary information was not retained
from one year to the next, other than what was printed in the ABR. Indeed
for years 1 & 2 all calculations were done manually on paper, for years 3 to
10 with limited storage capacity (or maybe for other reasons), running the
output routine obliterated the results of the prior year. Because I was
only looking at annual patterns and not total abundance over the
years, I arbitrarily (and as it became clear years later, wrongly) set the total
abundance of the species as equal each year. Knowing what the abundance estimate
or even number of records was, in those early years was no easy task, as the
system changed in the middle. I believe the absence of these graphs in recent
years hugely diminishes the value of the recent ABR in showing results for the
CURRENT year.
So it was to a large extent this exercise
and this species and what I saw was the potential and clearly the limitations of
what existed till then, which only output summary statistics separately by
species and by year with no connection of any kind between years, locations etc.
that lead me to creating the GBS database. Up till then there was no way of
knowing what species had been recorded in the GBS over all the years, comparing
any year to another apart from the previous one or even a coherent list of
names or addresses of contributors other than for each year separately or
anything like that.
So whilst the graph presented is
instructive, I would suggest against
analysing it beyond its use (with cautions partly raised by another recent sad
example - but no more of that here).
Philip
The number observed is a simple sum of
the number observed - abundance data by month is not something I have readily
available yet. Attached is the number observed data broken down into 5
year periods which may help with the question of changes in the
pattern.
Steve
| 