Monitoring Change on Mt. Diablo: 1 Year In
One year of growth and dieback from #morganfire02. Photographers from top left to bottom right: 2014-04-13 by Kris Kaufman, 2014-06-15 by Aleksander Gargenta, 2014-08-18 by Mark McElligott, 2015-01-04 by Sally.
It’s been just over a year since we installed our first camera stands on Mt. Diablo as a part of our Monitoring Change project, so we figured it was time to summarize. TL;DR: Even though almost no one took photos from the stands on repeat trips, we still got about 500 usable photos that show some interesting plant behavior! Yes, plants have behavior, they just behave very slowly, which is why timelapse photography is pretty much the only way to observe it. Without further ado, here are the movies (you'll want to turn on the HD):
Both crappy and cool, right? Crappy because, well, with varying cameras, weather, resolution, and cropping, it’s almost impossible to get the kind of consistency we’d need to make really smooth, David Attenboroughesque timelapses, and the computer we’re using to perform the image alignment (Ken-ichi’s aged MacBook) doesn’t exactly have the muscle to take full advantage of Petr’s sweet OpenCV-based alignment software, so it looks a bit choppy. But cool because look at what we’re seeing: at #morganfire01 you can watch the goldenbush (Ericameria linearifolia) go from buds to flowers to seed heads. In #morganfire02 you can see how the scorched earth bursts into greenery and flowers in May, to be succeeded by the perennial chamise stump-sprouting back to nearly a third of its former height in less than a year! In #morganfire03 you can compare what the grassland looked like after the fire in spring of 2014 (bare) to how it looked without fire in 2015 (lots of leftover grass from last year). And in #morganfire04… well, not much is going on there, since it was mostly chosen for its accessibility and not potential for biologically interesting stuff, but it is cool to see how the grass’s greening and browning seems temporally synced up with #morganfire01 and #morganfire03, which is frankly a lot easier to see if you view all four sites together:
Again, aesthetically jarring, but look at how the grasslands brown out just as the chaparral is greening up, and how the tarweeds in the summit grasslands of #morganfire03 bloom when almost everything else is toast. It’s pretty cool!
Our collaborators at California State Parks also had this to say:
As park managers at Mt. Diablo State Park, this is useful for park management. After a wildfire, there's often a lot of talk about the "devastation" of the fire. While some fires certainly can be ecologically devastating, fires can be beneficial for ecosystems. It's often a challenge to convey that message to visitors though, when the immediate aftermath of a fire is a charred landscape. Having videos like this, that show how quickly burned areas can bounce back, can be a great tool for park managers to use in communicating the benefits of fire in our parks. The videos also document the rate of re-growth of some plants, like the chamise in #morganfire02, what species bloom in what areas, and could document rare plants that park staff might not have records of yet.
So anyway, that’s what we’ve got. It’ll be really interesting to see if the timing is different this season. Will the whispering bells in #morganfore02 come back, and will they be as dense? Will the green last longer or end sooner? Will anything new bloom?
We’d like to give a big THANK YOU to everyone who’s participated in this project by taking photos, and we’d love to hear what you think. If you see anything else notable in these movies, please comment / tweet / send smoke signals (from a controlled fire!) / etc. If you think you can make a better movie, the data is all in a Google Spreadsheet and the software we use to fetch it and work with it is on GitHub, so have at it, and let us know what you make!
Also a huge thanks to our partners at California State Parks for being willing to do this and jumping through the many, many hoops that exist if you want to put some metal sticks in the ground in a state park.
The lessons we learned about this technique are also pretty interesting. As we mentioned, almost no one who took a picture from a stand went back to take more on subsequent trips. Maybe people don’t visit Mt. Diablo more than once a year, or maybe they just assumed taking one photo was enough. The variation in image quality was also pretty big, but we sort of expected that. More interesting is that you can use software to get around the alignment problem.
Using social media to facilitate image sharing worked out well enough. We hit some challenges when a pic of one of our signs (not taken from the sign) using one of our hashtags went semi-viral on Twitter, so we had to modify our data aggregation script to ignore retweets… which actually didn’t solve the problem, because a lot of people who retweet on Twitter don’t actually retweet, they just cut and paste, so we had to build in some moderation to our process, which ended up not being too much work. Spammers and tag hijackers also started using our tags a bit, but again, not too much to cause a moderation headache.
Using cloud services like Heroku (to host our automated aggregation script) and Google Sheets (to store and publish the data) worked pretty well, though the technology definitely has that kind of duct-taped DIY feel to it. You could imagine buttoning up the whole data side of the project into a web app, but then you’d have to maintain it…
The physical signs have lasted a full year. Ironically, someone spraypainted a tag on one of them, and on others the metal sign fell off the wood backing. The t-posts appear to be sturdy enough in the rocky soils of Mt. Diablo.
We’re glad to see others running with this idea and we wish them well. To anyone else who’s interested in giving it a try, we’ve done our best to document our methods in this instructable.
