Playing with fires.
Wildfires are a hot topic right now. Last year it was California, Australia and Brazil making headlines. Bolivia too was struggling to slow the spread of forest fires across its Protected Areas. This summer we’ve seen huge swaths of forest lost in Arctic Siberia as well as record-breaking fires scarily close to urban zones in Arizona.
With global temperatures rising each year and precipitation patterns becoming more extreme, it’s clear that the monitoring of fires will play an increasingly important role in protecting our forests.
So to that end, Global Forest Watch (GFW) has released a suite of new widgets to give users a sharper set of tools to monitor and analyse fires as they happen. I want to walk you through the new capabilities and show you some of the things you can do with them. Let’s take a look at some specific examples that are fresh in our memories.
Follow the smoke trail.
I’ve seen in the news that Arizona has had some historically big spikes this summer. Let’s start on the map and see if we can confirm that.
On the map, the raster VIIRS Fire Alerts source has been swapped for a vector tile cache so that you can view — and animate — more data in one go. This has the advantage of reducing the size of the tiles, improving performance, and allowing smoother rendering. Most importantly, this means we can put more data on the map at once! Whereas before you could view only the last seven days of fire events you can now view a whole fire season. Up to three months of data at a time.
So even if you’re reading an article that’s almost a month old, you can still check what happened then, what’s happened since, and what’s happening now.
You can also interact with individual alerts to see the characteristics of the fire event: a fire’s brightness, radiated power, and exact time and date of recording.
Since we know there have been wildfires in Arizona, let’s zoom in, take a look, and run a map analysis.
Looking closer at the hotspot just north of Tucson we can activate the Protected Areas layer and see that many of the June fires occurred inside the boundaries of Pusch Ridge nature reserve. These fires were triggered by lightning strikes, and exacerbated by high winds and abnormally dry conditions due to droughts in June.
Zooming back out we can jump to the country dashboards to compare States across the USA, using a few of our new widgets.
First, let’s see how Arizona ranks against other states in the last three months.
The significance ranking gives an idea of how many fire events a region typically experiences in some period of time. This takes the last eight years of data in the case of VIIRS in order to give some idea of what normal is by factoring in seasonal effects. You can account for up to 18 years if you switch to the MODIS dataset — which you can for any widget — swapping out high resolution for better detection performance in cloudy regions.
Simply put, a number between 1 and -1 means the region is experiencing a normal level of fires for that time period. Whilst a number above 1 can be considered more than normal (bad), a number less than -1 is less than normal (good). Framed this way, Arizona has had an unusually high number of fires!
The significance measure is good at a glance — it gives you a quick indication of how good or bad things are compared to usual. But it doesn’t paint the whole picture when used alone. As we can see North Dakota is the most significant, albeit with only 355 fires detected, suggesting that the state doesn’t usually see that many fires during these months.
In our case we can enrich that data by also looking at fire alert density — a measure of how intense, or close together (spatially) the occurrence of fires is. Arizona had, by far, the highest density with around 10 alerts per mega hectare, the total making up over a third of all fires in the USA at that time!
That gives us an idea that something unprecedented may be happening in Arizona. Next, let’s look at another new widget which gives us a clear picture of what’s happening compared to past data: the Cumulative Fire alerts widget.
Using this widget we can see the unprecedented nature of the Arizona fires. The year-to-date fires total is already way above expected levels for the entire year. We can also compare to previous years. The number of fires (so far) in 2020 is greater than the 2017 total — the previous worst year on record.
If you prefer to see the historical data laid out plainly instead of interpreting it with statistical bands, you can jump to the Historical Alerts widget, which has all historical fires at daily resolution.
Zooming way out on the horizontal scale here shows just how big these wildfires were, the 2020 flash dwarfing the peak fires of previous years.
Finding Fires.
Next, let’s take a look at another big hitter in the news — the Siberian wildfires. This time, though, let’s do it in reverse. Last time, I knew ahead of time exactly where to look: the fires were in Arizona, close to Tucson. However Siberia is a BIG place, and my knowledge of Russian geography isn’t great. So let’s use the Russia dashboards to guide us to some places of interest.
Jumping into the country dashboard again, let’s look at the region rankings for Russia to see where to look.
We can see that the district of Sakha ranks first in terms of the total number of fires as well as the fire density, and is in the top five for the significance of that fire count.
Looking at the accompanying map, it looks like there are many hotspots in the north-east of the region — which is strange considering the (usually) low summer temperatures at those latitudes.
Before jumping back to the map, I want to go one last level deeper and see where fires are the most intense. Let’s go back and check the ranking for locations within the Sakha Region.
So, using the dashboards, we seem to have found what we were after: a localised region where wildfires were rampant in the last few months, without any prior knowledge of the exact location!
Finally, lets jump back to the map and see if we can use some of the other tools offered by GFW to gather further details.
Again, we can see those huge (~1000 square kilometers) hotspots and I’m actually interested in checking one of them out. We know when the spike occurs (~mid June) so let’s zoom in and animate it.
We can see how fiercely that hotspot burned across the region, radiating massive amounts of heat and releasing carbon dioxide into the atmosphere.
Confirming the data.
I want to do one last thing: confirm that there were actual fires happening at that time so let’s activate the Recent Satellite service and access Sentinel-2 and Landsat-8 images of the location. Visual proof beats fake news, right?
Keep a closer eye.
I think I’d like to keep an eye on this little patch of land so I’ll add it to my custom Areas of Interest in my own ‘My GFW’ page — maybe with warmer temperatures and a drier climate these fires will become a yearly occurrence in this area?
Climate change is driving a higher occurrence of wildfires globally which is putting not only our forests at risk, but also the people who live near them and whose livelihoods depend on them. Fires also pose a health risk, with smoke and soot reducing the air quality and affecting communities for miles around. Therefore it’s important to monitor the areas that are beginning to emerge as wildfire risks more closely.
Well, we’ve taken a whistlestop tour of the new fires monitoring tools available in GFW. Now you can put them to good use throughout the 2020 fires season and watch how things evolve!
If you’re gonna play with fire (widgets) — let’s at least do some good with it.
AJ is a Scientist who loves space and skateboards. He finds ways to add context to data visualisations, making them easier to understand and use.
