Quest essay by Pasi Autio 14.7.2020
Figure: Natural forest fire in Russia.
Northern hemisphere summer – the season when forest fires in Siberia are on the loop. And usually every single new article about the Siberian forest fires somehow links them to climate change. Therefore it is good time to see how the forest fires has changed during the years. Is there really an increasing trend of Siberia forest fires as the news suggests and what is continuously predicted based on climate models?
With an area of 13.1 million square kilometres (5,100,000 sq mi), Siberia accounts for 77% of Russia’s land area. Majority of the Siberia is sparsely inhabited wilderness with little or no roads. Therefore, what sets on fire, usually burns until rain or other natural factor ends the fire. Southern Siberia also has extensive logging.
Getting reliable fire area data based on available literature seems to be problematic. According to the literature (1) USSR-era fire area data is unreliable and was consistently and severely underreporting fires on sparsely populated areas due to incomplete reporting structure that left most of the country unmonitored (6). The situation was improved only after western satellite data was taken in use by post-USSR Russia. But considering the size of Siberia and the fact that it is very sparsely populated, it is not wonder that no reliable data can be generated without the help of satellites. But even on satellite era some smaller fires goes undetected due to cloud cover or sensor detection limits (6).
After extensive literature study, I found no actual study providing satellite-based dataset for Siberian forest fires for post-USSR era either, which is strange considering how much coverage the Siberian forest fires have got lately. There seem to be an effort going on to create such a dataset for USSR-era years, however, by digitizing old satellite images taken since 1979, but let’s discuss that later a bit more.
Annual burned area in Siberia 1997-2016
Earlier essay I published (topic was Australian bushfires) made use of satellite data based on Giglio et al 2013. Giglio’s paper describes a fourth generation Global Fire Emissions Database (GFED4). This data set combines satellite records like the 500m MODIS burned area maps with active fire data from the Tropical Rainfall Measuring Mission (TRMM) Visible and Infrared Scanner (VIRS) and the Along-Track Scanning Radiometer (ATSR) family of sensors. This is excellent source to create a dataset also for Siberian fires.
The data is available at globalfiredata.org. Site provides a great analysis tool and of course the data itself, if you want to analyze it further. Currently the dataset provides burned area data for the years 1997-2016. It’s possible to select a region or country and choose several options about the source data from emissions to burned area (among others). The Analysis tools section also gives ability to use custom area shape definition. And that’s what I use to create my Siberia forest fire area dataset.
Figure: Definition of Siberia in use with the presented dataset
The area shapefile used tries to mimic greatest extent of Siberia area definition as closely as possible.
Figure: Siberia annual burned area km2
Figure provides the total burned area in Siberia for each year between 1997 and 2016 in km2. We can see that for last 20 years the burned area trend for Siberia is slightly downwards. No evidence whatsoever can be seen for fires getting any worse. The average burned area annually during that period was ~ 91181 km2 – about the size of Maine.
Years 1998, 2003, 2008 and 2012 were the four most severe fire seasons during this period. In general the inter-annual variability is great (3) with up to 4x differences between years. Forest fires during 2003 were ~ 203288 km2.
USSR-era burned area data
While studying the available literature I found that Stocks and Cahoon had started (ca 2010) a project (3) to digitize old AVHRR satellite images from the period 1979-2000 to build a satellite-based fire-area product for Siberia. This seems to have taken a while since the results of this work has nowhere to be found. The former student of Cahoon, Soja A.J., seems to have continued with this work (4) with Cahoon and Stocks and has presented the results in several conferences during 2018 and 2019. I asked the author whether the data is available for public consumption, but the according to the answer the data is still under validation.
The data is based on different (less sophisticated) instruments, algorithms and methods than GFED4 dataset presented above. Therefore it makes no sense to compare these datasets directly. But for purposes of finding out whether the Siberia forest fires have been getting worse, comparing the trends is interesting. Also, the datasets contain four overlapping years (1997-2000) and using these as a reference we can conclude that burned area on years 1985 and 1987 exceed 1998 and are among the most severe seasons during the satellite era.
The data presented in the conference (4) shows no increasing trend for Siberia burned area either.
In summary when we combine the AVHRR and GFED4 datasets we have 37 years (1979-2016) of burned area data for Siberia. During that time no increasing trend for the forest fires and no detectable signal for “climate change” can be found.
2020 Season fires
As usual, the news outlets are providing worrying stories about the forest fires in Siberia for this season. Greenpeace Russia has provided this piece of information (7):
“Greenpeace Russia’s forest programme, which analyses satellite data, said Saturday that a total of 9.26 million hectares—greater than the size of Portugal—have been impacted by wildfires since the beginning of 2020.”
Sounds bad. But how does this 9.26 million hectares (92600 km2) compare to earlier years? Once again, globalfiredata.org Analysis tool provides us this information. Cumulative burned area for Siberia from January to the end of June for selected years in the past:
- 2003: 15.4 Mha (154205 km2)
- 2008: 15.5 Mha (155114 km2)
If data provided by Greenpeace is correct (no source to verify it), the start of the 2020 fire season in Siberia has been one of the worst since 1997, but in no means the record.
- Siberian forest fires are extensive every summer with up to ~4x variations between the years
- Average burned area for the Siberia is ~ 91000 km2 / 9.1 Mha / 35200 sq miles – about the size of Portugal
- Contrary to the climate model predictions, no increased burned area can be found during 1979-2016 for Siberia
Russia’s Forests Dominating Forest Types and Their Canopy Density:
- Giglio, L., J. T. Randerson, and G. R. van der Werf (2013), Analysis of daily, monthly, and annual burned area using the fourth-generation global fire emissions database (GFED4),J. Geophys. Res. Biogeosci.,118, 317–328, doi:10.1002/jgrg.20042.
- Giglio, L., Boschetti, L., Roy, D.P., Humber, M.L., Justice, C.O., 2018. The collection 6 MODIS burned area mapping algorithm and product. Remote Sens. Environ. 217,72–85. https://doi.org/10.1016/j.rse.2018.08.005.
- Stocks, Cahoon 2010; Reconstructing Post-1979 Forest Fire Activity and Area Burned in Russia: NOAA AVHRR Analysis https://www.researchgate.net/publication/253580597_Reconstructing_Post1979_Forest_Fire_Activity_and_Area_Burned_in_Russia_NOAA_AVHRR_Analysis_Invited
- Historic AVHRR-derived Burned Area product and validation for Siberia (1979 – 2000) https://ui.adsabs.harvard.edu/abs/2019AGUFMGC24C..07S/abstract
- Vegetation fires and global change; White paper directed to UN
- Nearly 300 wildfires in Siberia amid record warm weather https://phys.org/news/2020-07-wildfires-siberia-weather.html
Siberia burned area dataset generated from globalfiredata.org:
Annual Burned Area: Area 1 (units: km^2)
area, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016
Area 1, 72515.49, 156759.94, 64566.03, 71509.61, 81478.54, 91947.5, 203288.34, 48805.25, 60614.74, 102303.24, 53793.62, 183118.31, 61870.23, 67793.46, 76604.99, 124019.83, 48602.92, 95504.16, 74647.34, 83883.75
Siberia shapefile used the generate the dataset above:
You can upload this shapefile to globalfiredata.org analysis tool to replicate the results in this essay.