Floating Forests and Google Maps
Often when citizen scientists view an image, they want context. Where is this? Am I really seeing kelp, or is this sand or mudflats? Fortunately, we have you covered. In the video below, I show you how you can view the metadata about each individual image, including how to view the area pictures in Google Maps. Now, the map on Google Maps isn’t going to be from the same time as the Landsat image, so, there may or may not be kelp in the same places. But you can at least get a better high resolution view of the area to make decisions about your classifications if you want it.
Does Citizen Science Consensus Alter Time Series?
One question that has come up a few times with our consensus classifications is, does the level really matter when it comes to looking at change in kelp forests over the long-term?
While our data isn’t quite up to looking at large-scale timeseries yet (we’re still digging through a few thorny methodological issues), I grabbed the complete dataset we have for the Landsat scene around Los Angeles from work Isaac is doing and decided to take a look. Note, this is totally raw and I haven’t done any quality control checking on it, but the lessons are so clear I wanted to share them with you, our citizen scientists.
After aggregating to quarterly data and then selecting the scene that had the highest kelp abundance for that quarter (i.e., probably the fewest clouds obscuring the view), we can see a few things. First, yeah, 1 classifier along is never good.
Note, I haven’t transformed the data to area covered, instead we’re just going with number of pixels (1 pixel = 30 x 30m). But, wow, we need consensus!
But what if we impose a more conservative filter? Say, a 4-10 user agreement threshold? What does that show us?
What I find remarkable about this image is that while we see the effect of decreases in detection when more and more citizen scientists have to agree on a pixel, the trends remain the same. This means that while we will try to chose the best threshold that will give us the closest true estimate of kelp area, that there will be multiple intermediate thresholds that give us the same qualitative results to any future attempts at asking questions of this data set.
This is a huge relief. It means that, as long as we stay within a window where we are comfortable with consensus values, this data set is going to be able to tell us a lot about kelp around the world. It means that citizen science with consensus classifications is robust to even some of the choices we’re going to have to make as we move forward with this data.
It means you all have done some amazing work here! And we can be incredibly confident in how it will help us learn more about the future of our world’s Floating Forests!
Lining Up Consensus
In working on a recent submission for a renewal grant to the NASA Citizen Science program, I whipped up a quick script that takes the data posted and overlays it with the actual image. I really like the results, so here’s one. Feel free to grab the script, data, and play along at home!
Kelpy Consensus
At Floating Forests, we get confronted with two issues a lot. First, how good is the data? A lot of scientists are still skeptical of citizen science (wrongly!) Second, many citizen scientists worry that they need to achieve a level of accuracy that would require near-pixel levels of zooming. We approach both of these issues with consensus classification – namely, that every image is seen by 15 people as soon as it’s noted that it has kelp in it. We can then build up a picture of kelp forests where, for each pixel, we note how many users select it as, indeed having kelp. You can read an initial entry about this here.
So, how does this pan out in the data I posted a few days back? Let’s explore, and I’ll link to code in our github repo for any who want to play along at home if you’re using R – I’d love to see things generated in other platforms!
The data is a series of spatial polygons, each one representing one level of consensus. After loading the data we can look at a single subject to see what consensus does.
I love this, as you can see both at the 1 threshold at least one person was super generous in selecting kelp. However, at the 10 threshold (unclear why we’re maxing at 10 here – likely nothing was higher!) super picky classifiers end up conflicting with each other so it looks like there’s no kelp here.
How does this play out over the entire coast? Let’s take a look with some animations. First, the big kahuna – the whole coast! Here are all of the classifications from 2008-2012 overlapping (I’ll cover timeseries another time).
I love this, because you can really see how crazy things are at a single user, but then they tamp down very fast. You can also see, given that we have the whole coastline, how, well, coastal kelp is! Relative to the entire state, the polygons are not very large. It’s a bit hard to see.
Let’s look at the coastline north of San Francisco Bay, from Tomales to roughly Point Arena.
Definitely clearer. You can also see we accidentally served up some lake photos, and folk probably circled plankton blooms. Oops! Now the question becomes, what is the right upper threshold? Time (and some ongoing analysis which is suggesting somewhere between 6-8) will tell!
If you have ideas for other visualizations you want to see, queries for the data that you want us to make, or more, let us know in the comments! If you want to see some other visualizations we’ve been whipping up, see here!
Floating Forests 
Recent Comments