In response to the recent attention on Jerseybirds, I figured I’d post a quick note.
You can read the original posts here:
https://lists.princeton.edu/cgi-bin/wa?A2=ind1010&L=Jerseybi&D=1&T=0&O=D&P=4594
https://lists.princeton.edu/cgi-bin/wa?A2=ind1010&L=Jerseybi&D=1&T=0&O=D&P=5396
and
https://lists.princeton.edu/cgi-bin/wa?A2=ind1010&L=Jerseybi&D=1&T=0&O=D&P=6361
To clarify, with some time and effort, radar data can be analyzed to estimate absolute numbers of birds migrating overhead. These values are highly variable under certain weather conditions, and therefore most studies that use such data choose them subjectively to minimize contamination by precipitation, insects, etc. Furthermore, radar data “out of the box” needs to be post processed to correct for certain inherent qualities which bias the data. This is NOT what I do on woodcreeper.com (but it IS what I do for NJ Audubon). Most of what I do on woodcreeper.com is off-the-cuff analyses of the weather forecast, backed up (or not) by the nightly radar, and my (limited but growing) understanding of bird migration over New Jersey (and to a lesser extent, elsewhere). Trying to estimate bird density (on a very non-quantitative scale of “low” to “high” and sometimes “really fricking crazy”) every night has an inherent level of inaccuracy (sometimes very high!). Couple that with the fact that birds overhead does NOT mean birds on the ground the following morning, and you begin to understand that predicting birding conditions based on weather and radar is both an art AND a science (with art trumping science under conditions where the predictive properties of weather or radar decline). My post yesterday was tongue-in-cheek towards Mike Hiotis, but in total seriousness, what I do on a daily basis is meant to better understand what factors are driving birds into the places we find them (and don’t find them). So any observations to that end are productive towards our general understanding of the relationships between weather, what we see on the radar, and birds on the ground.
If I “say” that there are no birds migrating, it’s almost always based on the radar (read: the radar detected no birds migrating). West winds last night could easily have brought an influx of birds from PA, but they weren’t showing up on the radar last night- so the only way to know would have been to a) listen during the night for flight calls and/or b) get out and bird. Walt should not feel compelled to qualify his report- he went birding, there were new birds = some birds migrated* (*assuming he or someone else was there yesterday, and counted less birds). That’s interesting! and also gives us one more data point where the radar just couldn’t pick it up.
Good Birding
David
4 responses to “Recent posts on Jerseybirds”
I think that like many other birders, I use your site, posts, tweets as major factors in choosing to get out early (though I probably would anyway) and check out migration and birds on the ground. I’m not sure you get enough ‘ground-truthing’ as you say, to validate your prognoses, other than perhaps Cape May, and a few spots that are regularly visited by those who comment – Palmyra/Delaware River, Liberty SP, Sandy Hook, Chimney Rock etc.
How can we help improve the science part?
As Susan commented, what do you think has changed/improved since you’ve been working on radar?
Do you track ‘success’ of radar/ground prediction/outcome – I would say you’re about 75% good, although as you righly say, migration shown on radar at night does not equal birds on the ground, but should be a good indicator.
I’m lucky that I can and do stop regularly at LSP and Chimney Rock, which are my primary local birding ‘patches’ along with the Great Swamp, and I’m happy to report..
As a Brit, I can state that we in NJ are incredibly lucky to witness the migration experienced here, where others around the world have to work much harder to find the numbers of birds that are readily accessible to us all. Having some extra science and enthusiasm to help the quest is icing on the cake.
Now, off to the watch repairer…….
Ugh… I just wrote several pages in response to your great comments, only to lose it when I tried to save a draft (comments don’t save in draft form! doh!)
So, to summarize: yes, I’ve gotten much better at predicting birding conditions on the ground over the years. Basically this boils down to a better understanding of the temporal variation in migration (who migrates when, and in what relative numbers) and a better understanding of the weather that drives migration over New Jersey (and especially how this varies over space and time). Coupled with that is an improved understanding of how the radar portrays such events.
There are gaps in my understanding for sure, but as Simon points out the lack of ground trotting is probably the main obstacle. I have never visited the NP Dredge Spoils nor Palmyra, but I think I have a decent handle on both of them in terms of good migration conditions simply because of the consistent reporting from those locations (much gratitude owed to Sandra Keller). I can’t say the same for the Meadowlands or the Great Swamp. For all of the crack birders living in Cape May, it can still surprise the seasoned veteran, which is a humbling experience for someone who just recently arrived. Still, Cape May is a fall hotspot due to cyclical weather patterns and geographical location, just as is Sandy Hook, and Garret Mountain in the spring. Thus predicting good conditions at these locations is much easier when the weather follows typical patterns. In those cases, the radar is simply a confirmation of migration density and direction.
In order to make woodcreeper.com more scientific we would need consistent ground-based observations at locations we expect to be good, and expect not to be good, based on the radar and weather. On my end, I would need the time and support to build predictive migration models based on many years of archived radar data and weather measurements. The latter part is ongoing, and all of the work I’m doing now should eventually lead to that end. The former is something that has been started by the consistent posters to this site, but which can be greatly augmented through other data sources such as eBird. In fact, my goal is to develop spatially explicit predictive models which would assign probabilities to actual eBird hotspots (“there is a 65% chance of increased bird density at xxx location tomorrow”). The underlying theme here, though, is the conservation of migratory species. If we can finally quantify migration across large spatial scales, we can then begin to estimate national and global populations and prioritize those needing immediate conservation action.
Okay, it’s past my bedtime, and the Higbee Dike is going to be teeming with birds tomorrow morning, so I’m going to sign off. See you back at 5:30am, and then maybe in the field. Thanks again for your continued support, and for those of you posting comments, you make this site what it is.
Cheers
David
David,
I think we are fortunate that you help us all out with your migration forecasts. Thanks!
David,
Your work here on Woodcreeper inspired some of my thinking on nexrad radar and birding. While I’m not as academically interested in improving the scientific accuracy of the interpretation you do, I think the education you’re doing in this post is related to my work. Quickly, just how birding changes when too much emphasis is placed on migration forecasts.
I’ve just posted something of a summary on my blog if you’re interested in reading a bit more.