application of geoinformatics in forestry · wildfires of siberia under climate change the current...

11TH Session of the Winter Forest School

«Application of geoinformatics in forestry»

Forest Research Institute, Sękocin Stary, 12-14 March 2019

Remote sensing

in monitoring wildfires in Siberia

Evgenii Ponomarev

Federal Research Center «Krasnoyarsk Science Center SB RAS»V.N. Sukachev Institute of Forest, SB RAS

Siberian Federal UniversityKrasnoyarsk, Russia

____________________________________________________________________________________________________

Actualities

1/21

I.Wildfire impact is the main natural disturbances in forests of Russia and mostly in Siberian boreal zone.

In Siberia the annual area burned makes up to 17 million ha.

The increasing trend of forest burning is evaluated over recent years. (Conard et al. 2002; Soja et al. 2004; de Groot et al., 2013; Ponomarev, Kharuk, Ranson, 2016).

II.

There are 4 zones of forest fire monitoring in Russia today.

According to them more than 50% of forests in Russia is under remote monitoring only. Satellite imagery and remote sensing estimations are the only available data on

wildfire activity there.

Satellite monitoring allows to obtain real-time data on forest fires and provides with attributive information on all detected active burnings and their daily dynamics.

11TH Session of the Winter Forest School «Application of geoinformatics in forestry» Forest Research Institute, Sękocin Stary, 12-14 March 2019

Territory zoning in terms of wildfire monitoring and preventing

A) Levels of Natural fire danger

I – ground-based fire monitoring (7%),

II – aerial monitoring (42%),

III – satellite monitoring and selective fire extinguishing (20%),

IV – satellite monitoring and fire extinguishing in cases of settlement threats (31%)

V – protected areas (State reserves, National parks etc.)

A

C) Percentage of forest cover


B

B) Anthropogenic impact levels D) Types of preventing monitoring

C

D

2/21

Tasks for the system of remote fire monitoring

Fire Weather condition analysis and weather fire danger estimation;

Thermal anomalies detecting and wildfire attributive data collecting;

Real-time mapping of fire activity and dynamics.

Long-term wildfire database preprocessing and uploading in GIS;

Geospatial analysis of wildfire, current trends and prognoses;

Assessment of fire impact on the scale of ecosystems / post-fire forest monitoring;

Direct carbon emissions estimating.

Daily

real-time

monitoring

Analysis

of fire ecology

and post-fire

effects


3/21


4/21

Remote sensing data on wildfires in Russian Federal agencies

«Kaskad» (Cascade)

Emergency Situations Monitoring agency-level portal of EMERCOMof Russia

Supported since 2009 by the Regional department in Krasnoyarsk

Reduced version is available at: http://space.akadem.ru/int/

«ISDM - Rosleskhoz»

Remote Monitoring Information Systemof Federal Forestry Agency

Available at: http://public.aviales.ru/main_pages/public.shtml

Estimating the wildfire energy characteristics, intensity of burning, crown burning

Monitoring data on wildfires for science

Spatio-temporal characteristics of wildfire frequency (fire weather, heat and moisture supply, natural fire danger)

Post-fire effects monitoring and prognosing (thermal anomalies, impact on thawing depth of the seasonal thawed layer, forest mortality/recovery)

1

2

3

4


5/21

GIS of Wildfire monitoring

V.N. Sukachev Institute of Forest, Siberian Branch of Russian Academy of Sciences

Available for 1996 – 2018

Daily information for every wildfire is available as GIS layer also.

The database contains over 2104 records per year and ~5106 in total for 24 years.

The main topics

Instrumental assessment of direct fire emissions of carbon based on real-time satellite data

II

I

III

Area of interestAgreement between The Global Fire Monitoring Center and

Sukachev Institute of Forest, 2017

IV

E u r a s i a


6/21

Central Siberia, Yenisey transect

AOI of Regional Eurasia Fire Monitoring Center /The Global Fire Monitoring Center

South Siberia Altai-Sayan Region

Satellite imagery for wildfires monitoring

Terra/MODIS2017

Active burning and post-fire pattern

Sentinel-22016

Landsat-8/OLI2017

AQUA/Modis SNPP/VIIRS

Raw data on thermal anomalies/burned areasAWildfire database in GISС

Pre-processing

B

D Analysis


7/21

GIS layer of long-term history of wildfires


8/21

Data pre-processing in GIS

- final fire polygon aggregating;- polygons of area burned re-calculating;- total fire scar area validating;- additional attributive information (region, forestry, forest type, nearest settlements etc.)

Probability of active burning detecting vs the area and target temperature

%

Active zone area, m2

0

200

400

600

800

1000

1200

1400

12 14 15 16 18 19 20 22Category of area

Num

ber

of fire

s

Categories of fires in terms of size:

(1) Wildfires up to 1000 ha;(2) «short-lived» objects/ agricultural burning;(3) large-scale wildfires

Categories of Wildfires in database


9/21

0

100

200

300

400

500

600

12 14 15 16 18 19 20 22

Категории

Количест

во п

ож

аров

а)

0

50

100

150

200

250

300

350

400

12 14 15 16 17 18 19 21 22

площадей

б)

A histogram of data on forest fire events: a) taking into account all detected thermal

anomalies, b) after the exclusion of «short-living»

objects registered during the spring

0

5

10

15

20

25

1996 2000 2004 2008 2012 2016

Год пожароопасного сезона

Числ

о п

ож

аров,

тыс.

1 2 3

а)

0

5

10

15

20

25

1996 2000 2004 2008 2012 2016

Год пожароопасного сезона

Общ

ая п

лощ

адь,

млн г

а 1 2 3

б)

Wildfire numbers, ths

Total area, MHa

Fire Radiative Power / burning intensity estimations

100

1000

10000

100000

1 5 9 1317 21 2529 3337 4145 49 53

Number of fire observation

FR

P, M

W

1

2

Fire radiative power dynamics

To estimate the heat radiation power from the zone of active burning the Terra/MODIS data in the range of 4 mkm / standard product MOD14 were employed(Kaufman et al., 1998; Justice et al., 2002).

The method was adapted for burning conditions in Siberia;The threshold technique was implemented to classify fires in term of burning intensity: low-, medium- and high-intensity including crown burning (probability ~60%).

The characteristics strongly correlated with a burnup rate and fire front velocity (initial data for modelling of scenario of a wildfire).


10/21

Extreme and crown wildfires in Siberia

Sporadic peaks of FRP corresponded to the extreme

burning phase

100

1000

10000

100000

1 8 15 22 29 36 43 50 57 64 71 78

Number of fire observation

FR

P, M

W

1

2

Annually extreme intencity burning zones were classifyed presented up to 10-13% of total area of Siberian wildfires.

Averaged estimation for crown burning is 5.51.2% of the total fire number, that is about 8.5% of total area burned annually.

FRP maxima

Thermal anomalies from the Terra/MODIS imagery of an active burning


Pixel and sub-pixel re-analysis1-km resolution image

11/21

Remote system for fire monitoring is the main source of data for analysis and forecasting of fire impact at the scale of ecosystems of Siberian forests and globally while the quota (contribution) of forest burning in Eurasia is strongly significant.

Wildfires of Siberia under climate change

The current dynamics of fire regimes in Siberia is determined by a complex of factors, such as temperature anomalies, change and redistribution of precipitation.

The first decade of the 21st century is characterized by an increase in frequency of fires and burned areas.

According to forecasts in the 21st century, direct wildfire emissions of carbon under current trend can rise from 120-140 Tg at present (Shvidenko et al., 2011) up to 230-240 Tg per year (Zamolodchikov et al., 2011).

4

5

6

7

8

9

10

11

1900 1920 1940 1960 1980 2000 2020

Tem

pera

ture

anom

aly

, °С

1901 - 1990

1990 - 2013

I

II

a)

-0,7

-0,2

0,3

0,8

1900 1920 1940 1960 1980 2000 2020

SP

EI

b)

Long-term trends for temperature and The Standardised

Precipitation-Evapotranspiration Index for Siberia


12/21

Heat and moisture supply (top) trends vs RBA (bottom)

Long-term averaged relative burned area (RBA) per year, %

Wildfires vs meteorological anomalies and climate trends

Relative burned area per year is in the range of 0.1% — 14.5%Averaged for Siberia — 1.19%, for western Canada — 0.56% (deGroot et al., 2013).

Fire season scenario

P{E} (min–max)Period, yeas

RBA, %(Min–Max)

I extreme 0.18–0.20 8 ± 3 4.5–14.5

IIa moderate/spring

0.24–0.57 4 ± 1 0.5–1.5

IIb moderate/summer

0.24–0.38 3 ± 1 1.0–4.0

III low 0.19–0.48 4 ± 2 0.01–0.3

Angara Evenkiya Yakutia


13/21

Assessment of fire impact and monitoring of post-fire changes could provide "traditional" methods for analyzing vegetation indices, as Normalized Difference Vegetation Index (NDVI);Normalized Burned Ratio differencing (dNBR) etc.

The real time classifying can be produced using the FRPtechnique data.

The approach can be used for estimating the initial fire impact and for the furtherprognosing of the post-fire vegetation cover state.

Vegetation cover classification for the post-fire scar:a) initial Landsat imageb) uncontrolled classificationc) NDVI classifyingd) dNBR classifying

e) Real-time classification using Fire Radiative Power data / intencity estimates

Fire impact classifying

a b c

d e

Fire impact “severity”


14/21

Post-fire effect monitoring

1-2 year

15 year

24 year

Larch forest recovery process monitoring (Evenkiya, Siberia)

Year after

the burning

NDVI

relative, %

Absolute temperature anomaly, °C

T

relative, %

1 53.510.7 6.57.2 47

5 21.07.8 3.84.9 30

10 9.05.0 3.44.6 20

Averaged characteristics of thermal and vegetation anomalies for Larch forests of permafrost zone


y2 = 41,7e– 0,08x

R2 = 0,81

y1 = 57,6e– 0,2x

R2 = 0,98

0

20

40

60

0 5 10Период после пожара

Ан

ом

ал

ия

ND

VI

отн ,

% .

0

10

20

30

40

50

Ан

ом

ал

ия

Tо

тн ,

% .

1

2

0

5

10

15

20

0,0 0,2 0,4 0,6Аномалия NDVIотн

Z,

%

1 2 3

r = 0,7

15/21

Modeling the fire impact

Z, % - estimations of depth of the seasonal thawed layer

rate of loss fromthermal anomalies

Fire intensity for emission calculating

The ratio of fire areas in terms of intensity

Fractiles of intensity:

I) < FRPmean–σ, II) от FRP mean–σ до FRP mean+σ,

III) > FRP mean +σ

σ - is standard deviation

The power of heat radiation is linearly related to the amount of burnt biomass (Wooster et al., 2002).

Fire Radiative Power data are the initial information for calculating the amount of emissions during different stages of each wildfire.

11.72

44.6043.67

0

25

50

75

I II IIIКвантиль FRP/Интенсивность

До

ля п

ло

ща

ди

, %

.

б)

11.68

46.0442.28

0

25

50

75

I II III

Пл

ощ

ад

ь,

% .

а)

43.64 42.92

13.44

0

25

50

75

I II III

До

ля п

ло

ща

ди

, %

.

г)

/FRP

10.94

41.74

47.32

0

25

50

75

I II III

Пл

ощ

ад

ь,

% .

в)

Квантиль интенсивности

Larch Pine

Dark Decid. /Mixed


Fire classifying for the territory of Siberia

16/21

Variation of the registered radiation power (FRP) in relation to fire parameters in model equations

Wildfire area (m2)

Combustion completeness ()

Fuel load (kg/m2)

Burned biomass (Seiler, Crutzen, 1980 )

Wildfire polygon classification / intensity category

Adopted estimations of burned amount of fuels (kg);Adopted method for direct carbon emission estimates (Tg С/year)

Evaluation of wildfire emissions

StandS, mlnha peryear

“Standard” season

Extreme season

% of emission

(min–max)

TgС/year

t С/ha

Tg С/year

t С/ha

Larch 2.765 42.9 15.5 52.0 18.8 51.6–62.4

Pine 0.656 11.0 16.7 11.8 18.0 13.2–14.2

Dark coniferous

0.153 1.9 20.4 3.1 20.4 2.3–3.7

Decidious/mixed

0.275 3.8 13.7 4.717.2

44.5–5.7

а) FRP vs burnup rate (kg/m2sec) for different sub-pixel active burning area: 1) 1000 m2, 2) 500 m2, 3) 250 m2;

b) FRP vs fire front velocity and fuel load(kg/m2): 1) for 1.5 kg/m2 and β = 0.55, 2) for1.5 kg/m2 and β = 0.4, 3) for 2.5 kg/m2 and β = 0.55, 4) for 0.7 kg/m2 and β = 0.4

0

50

100

150

200

0 0,05 0,1Front velocity, m/sec

FR

P,

МВ

т

1 2

3 4

b)

0

50

100

150

200

0 0,05 0,1 0,15

Burnup rate, kg/m2sec

FR

P,

MW

1

2

3

a)


17/21

Direct emission 83 ± 21 Tg С/year is less in comparison with assessment for Siberia 112 ± 25 Tg С/year, according to standard method (Soja et al., 2004).

Emission values per year vary from minima 20–40 Tg С/year (during 2004, 2005, 2007, 2009, 2010) up to 230 Tg С/year during extreme fire season of 2012.

The total emission statistics includes 33–37%, 47–49% and 14–17% from wildfires oflow-, moderate- and high intensity in terms of FRP. And specific emission values were8.7, 12.0 and 15.4 tons С/ha correspondingly.

Method

Burned biomass (М) Direct emissions (C) Mrel и Сrel

1012 kg σε

for

α= 0.1

Tg С / year

σ ε % σε

for α=

0.1

“Standard”

(Seiler, Crutzen,

1980 )

0.192 0.131 0.067 111.9 68.4 25.4

17.3 1.6 0.8Standart+

FRP

classification

of fires

0.159 0.108 0.055 83.1 56.5 21.0

Long-term emission estimates


18/21

Trend of the direct fire emissions in Siberia

Variations of direct carbon emissions from Siberian fires in the time interval 2002-2016:

a) trend based on the multi-year series (p <0.05); b) In relation with air temperature anomaliesfor Siberia

The updated estimates of direct fire emissions in Siberia during 2002–2016 were 83 ± 21 Tg C/year in average. Current linear trend (R2=0.56) was evaluated for the last 15 years.

A trend of significant increase in direct fire emissions is observed. According to the trend fire emissions in Siberia at the end of 21 century will rise up to 220, 700 and 2300 Tg C/year in RCP2.6, RCP4.0 and RCP8.5 scenarios, correspondingly.

1

2

R2 = 0.56

0

100

200

2002 2006 2010 2014

Year

С,

Tg/year

a)

y = 32.7exp(0.59x)

r = 0.51

10

100

1000

0,0 0,5 1,0 1,5 2,0 2,5Temperature anomaly, °С

b)


19/21

Conclusion


20/21

Satellite data and GIS technology is the important part of fire preventing and monitoring system for Siberia.

Real-time data on forest fires is available for different agency including the Federal forest service, EMERCOM and fire ecology science as well.

Precise database of wildfires in Siberia is available for 1996 – 2018 in GIS layer format. And number of technologies for fire characteristics assessment were adapted to the conditions of burning in boreal forests of Siberia at present time.

Remote system for fire monitoring is the main source of data for analysis and forecasting of fire impact at the scale of ecosystems of Siberian forests and globally since the assessment of forest burning in Eurasia is strongly significant.

Publications


Ponomarev E.I., Kharuk V.I. (2016) Wildfire Occurrence in Forests of the Altai–Sayan Regionunder Current Climate Changes // Contemporary Problems of Ecology. 2016. Vol. 9. № 1. P.29–36. DOI: 10.1134/S199542551601011X

Kharuk V.I., Ponomarev E.I. (2017) Spatiotemporal Characteristics of Wildfire Frequency andRelative Area Burned in Larch-dominated Forests of Central Siberia // Russian Journal ofEcology. Vol. 48, No 6, p. 507–512. DOI: 10.1134/S1067413617060042

Ponomarev E.I., Skorobogatova A.S., Ponomareva T.V. (2018) Wildfire Occurrence in Siberiaand Seasonal Variations in Heat and Moisture Supply // Russian Meteorology and Hydrology.Vol. 43, No. 7, p. 456–463. DOI: 10.3103/S1068373918070051.

Ponomarev E.I., Shvetsov E.G., Kharuk V.I. (2018) Intensity of wildfires for direct fire emissionsestimating // Russian J. Ecology. Vol. 49. №6. P. 492–499. DOI: 10.1134/S1067413618060097

Ponomarev E.I., Shvetsov E.G., Litvintsev K.Y., Bezkorovaynaya I.N., Ponomareva T.V.,Klimchenko A.V., Ponomarev O.I., Yakimov N.D., Panov A.V. (2018) Remote Sensing Data forCalibrated Assessment of Wildfire Emissions in Siberian Forests // Proceedings. 2018. Vol.2. №7 (348). 7 p. DOI: 10.3390/ecrs-2-05161.

Ponomarev E.I., Ivanov V.A., Korshunov N.A. (2014) System of Wildfire Monitoring inRussia (pp.187 - 205). In: D.Paton and J.F. Shroder (ed.) Wildfire Hazard, Risks andDisasters. Elsevier, 2015. – 284 p. ISBN: 978-0-12-410434-1. DOI: 10.1016/B978-0-12-410434-1.00010-5

20/21

http://dx.doi.org/10.1016/B978-0-12-410434-1.00010-5

Thank you!


Remote sensing

in monitoring wildfires in Siberia

Evgenii Ponomarev

Federal Research Center «Krasnoyarsk Science Center SB RAS»

V.N. Sukachev Institute of Forest, SB RAS

Siberian Federal University

Krasnoyarsk, Russia

application of geoinformatics in forestry · wildfires of siberia under climate change the current...

Documents