GOFC-WARN Regional Workshop

Remote Sensing Based Monitoring of Land Surface Dynamics in West Africa

MONITORING WILDFIRES AND THEIR DRIVERS

IN COTE D’IVOIRE
24-25 March 2021

Jean-Luc Kouassi 1,2*, Narcisse Wandan1 and Cheikh Mbow3

1 Institut National Polytechnique Félix Houphouët-Boigny (INP-HB) ;
2 African Global Impact Raiser (AGIR);
3 Future Africa, University of Pretoria, Pretoria, South Africa
 INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Background

Wildfires, a global ecological disaster (deforestation, damage and deaths)

Camp Fire, California Mati and Neos

- USA – Dec. 2018 Voutzas fires, Greece
83 dead, 62,000 ha and – July 2018
13,500 houses destroyed 99 dead, 172 injured,
entire city and 300
vehicles devastated
1 2

Black Saturday Wildfires,

Bushfire, Victoria - Côte d’Ivoire
Australia – Febr. 2009 1982/83 : + 100 dead, 1.8
173 dead, 38,500 ha, million ha and + 1000
2,000 houses and 6 huts burned
schools destroyed 2016 : 17 dead, 26,000
ha and 200 huts
3 4 destroyed
Fig. 1 : Wildfire activities around the world 2
INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Study area
West Africa

Côte d’Ivoire

Fig.2 : Active fires in Dec. 2005

Active fire in Central Cote d’Ivoire
Area: 322 462 km2
Longitude (2°20′ W and 8°40′ W) and latitude (4°20′ N
and 10°50′ N)
A tropical climate of Aw type (warm and humid) with 3
subclimates (Guinean climate, Baoulean and Sudanian)
Fig. 3: Map of Cote d’Ivoire
Ecosystems: Forest ecosystems (south), mosaic
forest‒savannah ecosystems (centre), savannah
ecosystems (north) and mountain ecosystems (west)
Population: 22.7 million inhabitants (RGPH 2014)
3
INTRODUCTION METHODOLOGY
  Challenges
Wildfires, a real recurring scourge
(1982/83, 2016, etc.) impacting
biodiversity, society and the economy
Climate change manifestations (rainfall
and hydrometric deficit, temperature
increase and droughts)
No fire risk management and early
warning system (evolution, factors and
forecast)
RESULTS PERSPECTIVES
  Objectives
Analyze the dynamics of wildfires and
determine the causes and factors
influencing their distribution in the
different phytogeographic zones of Côte
d'Ivoire.
 determine the regime and trends of
wildfires in Côte d'Ivoire;
 determine the impact of climatic factors
on the fire occurrence
 assess the impact of environmental
factors on the outbreak and distribution
of wildfires.
4
 INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Methodology

Objectives Material Methods

- MODIS FIRMS Active fires data (2001 -
- Trend analysis
2019)
- Descriptive analysis (fire counts and
Wildfire dynamics - MODIS Burned Area Product MCD45A1
burnt areas)
(2001 - 2019)
- Sofware : R and QGIS
- Household survey (3 villages) and 1
- Questionnaire focus group (3 men and 3 women) per
Local perception
- Software : R (questionr) village
- Bivariate analysis
- Climate variables from synoptic stations
- Cross-correlation (climate vs fire var.)
- Environmental data (vegetation,
Driver analysis - Chi-squared test (ecoregion)
topography, socioeconomic data)
- Kruskal-Wallis test (categories)
- Software: R

5
INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Retrospective dynamics of wildfires

Fire season: Fire season:
Nov - May Nov - March

Febr: 5,939 fires Jan: 14,044 fires

Jan: 1.31 million ha Jan: 3.16 million ha

Fire season:
Nov - March Fire season:
Nov - March

Dec: 56,540 fires

Dec: 49,620 fires
Dec: 28.20 million ha
Dec: 24.53 million ha

Kouassi et al. under review 6

Fig. 4: Monthly number of wildfires and burnt areas
INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Retrospective dynamics of wildfires

2007: 2,057 fires 2008: 2,838 fires
2007: 0.486 million ha 2016: 0.611 million ha

2005: 8,732 fires 2008: 12,127 fires

2005: 3.58 million ha 2005: 4.35 million ha

Kouassi et al. under review 7

Fig. 5: Annual number of wildfires and burnt areas
 INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Retrospective dynamics of wildfires

SZ: 104,858 fires SZ: 39.54 million ha

Mean: 5,519 fires/year Mean: 2.08 million ha/year

PFZ: 28,519 fires PFZ: 7.97 million ha

Mean: 1,501 fires/year Mean: 0.42 million ha/year

FZ: 21,471 fires FZ: 3.57 million ha

Mean: 1,130 fires/year Mean: 0.19 million ha/year

CI: 154,848 fires CI: 51.08 million ha

Mean: 8,150 fires/year Mean: 2.69 million ha/year

Kouassi et al.
under review
Fig. 6: Spatial distribution of active fires and burnt areas by ecoregion from 2001 to 2019 8
INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Retrospective dynamics of wildfires

Tab. 1 : Trends of the number of wilfires and burnt areas

Number of wildfires Burnt areas

Location
Sen's slope Test Z p-value Sen's slope Test Z p-value
Forest zone –0.4518 –10.278 p < 0.001 –0.2077 –4.865 p < 0.001
Pre-forest zone –0.3498 –8.026 p < 0.001 –0.1110 –2.673 p < 0.01
Sudanian zone –0.2792 –6.425 p < 0.001 –0.1025 –2.392 p < 0.05
Côte d'Ivoire –0.3411 –7.716 p < 0.001 –0.1409 –3.258 p < 0.01
Kouassi et al. under review

Significant downward trends of the number of wildfires and burnt areas in all the ecosystems

Consistent with findings of Dwomoh and Wimberley (2017) and N’Datchoh et al. (2015) in West Africa

Climate change mitigation efforts (awareness raising, reforestation, etc.)

9
 INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Impact of climate variables on wildfires

Tab. 2: Summary of the significant cross-correlation functions identified between the pyrological variables and
the climate variables
Fire CCFmax*[lag (month)]
Ecoregion
data PRCP MAX MIN TM TA DEWP RH SLP VPD VISIB MXSPD WDSP
Forest
–0.418 [–1] 0.626 [0] 0.486 [2] 0.619 [1] 0.634 [0] 0.374 [2] –0.744 [0] - 0.761 [0] –0.630 [–1] 0.245 [1] 0.282 [6]
zone
Pre-forest
Number –0.471 [–1] 0.625 [1] –0.488 [0] 0.638 [2] 0.703 [0] –0.830 [0] –0.781 [0] - 0.728 [0] –0.785 [0] –0.375 [–1] –0.316 [–1]
zone
of
Sudanian
wildfires –0.420 [–5] 0.364 [2] –0.458 [0] 0.499 [–9] 0.424 [0] –0.474 [0] –0.429 [0] –0.155 [3] 0.414 [1] –0.474 [0] 0.207 [3] 0.290 [3]
zone
Cote
–0.517 [0] 0.598 [1] 0.535 [3] 0.703 [2] 0.600 [0] –0.546 [0] –0.584 [0] - 0.631 [1] –0.743 [0] –0.323 [11] –0.380 [11]
d’Ivoire
Forest
–0.329 [0] 0.402 [2] 0.296 [2] 0.422 [2] 0.533 [0] –0.354 [0] –0.590 [0] –0.168 [3] 0.541 [0] –0.675 [0] 0.195 [2] –0.274 [–1]
zone
Pre-forest
–0.482 [0] 0.580 [1] –0.549 [0] 0.627 [2] 0.614 [0] –0.719 [0] –0.695 [1] 0.131 [7] 0.712 [1] –0.777 [0] –0.352 [0] –0.270 [–1]
Burnt zone
areas Sudanian
–0.437 [–4] 0.462 [–4] –0.536 [0] 0.564 [–9] 0.459 [0] –0.552 [1] –0.497 [1] –0.193 [–14] 0.455 [1] –0.500 [1] 0.197 [3] –0.283 [–2]
zone
Cote
–0.452 [0] 0.554 [2] 0.459 [3] 0.592 [2] 0.574 [1] –0.581 [1] –0.629 [1] –0.140 [–10] 0.611 [1] –0.657 [1] –0.295 [0] –0.332 [0]
d’Ivoire
PRCP: Precipitation, TM: Average temperature, MAX: Maximum temperature, MIN: Minimum temperature, TA: Thermal amplitude, WDSP: Average wind speed, Kouassi et al.
MXSPD: Maximum wind speed, RH: Average relative humidity, VPD: Vapour pressure deficit, SLP: Sea level pressure, DEWP: Dew point, VISIB: Average visibility, under review
CCFmax *: Maximum significant value (p <0.05) of the cross-correlation function.
10
 INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Impact of environmental variables on wildfires

Table 3. Cross-table of wildfire occurrence per category of environmental drivers.
Pre-forest Sudanian Pre-forest Sudanian
Environmen Forest zone Côte d'Ivoire Environment Forest zone Côte d'Ivoire
Categories zone zone Categories zone zone
tal variables al variables
% % % % % % % %
[0-100[ 17.26 0.93 0 2.56 [0-1[ 8.18 9.44 2.04 4.26
Elevation Distance to [1-5[ 72.2 74.86 35.33 47.72
[100-300[ 65.29 65.62 39.4 47.82
(m) localities (km) [5-10[ 18.15 13.1 29.33 24.79
> 300 17.45 33.45 60.6 49.62
> 10 1.48 2.6 33.29 23.23
[0-5[ 39.81 47.14 56.81 52.86
[0-1[ 33.8 30.97 9.94 17.12
Slope (%) [5-10[ 37.31 38.56 35.01 35.98 Distance to
[1-5[ 54.64 53.08 35.32 41.27
> 10 22.88 13.31 8.18 11.17 roads and
[5-10[ 10.46 12.43 23 19.32
Acrisols 14.67 18.23 36.06 29.81 railways (km)
> 10 1.1 3.52 31.74 22.29
Alisols 11.69 7.48 17.09 14.57 [0-1[ 10.39 5.44 9.72 9.03
Cambisols 12.77 7.84 6.66 7.72 Distance to
Soil type [1-5[ 28.48 18.52 28.3 26.52
Luvisols 0 0 10.06 6.82 waterways
[5-10[ 26.38 22.53 24.37 24.31
(km)
Plinthosols 51.5 62.38 23.78 34.73 > 10 34.75 53.52 37.62 40.15
Other soils 18.71 8.12 12.59 12.62 Population [0-25[ 46.96 66.35 70.27 66.32
Croplands 2.47 0.05 0.14 0.45 density [25-75[ 15.01 7.14 3.51 5.78
Forests 1.58 0.21 0.11 0.33 (pers./km2) > 75 38.03 26.51 26.22 27.91
Landcover Grasslands 0.52 1.06 19.7 13.61 [0-1[ 5.01 3.36 2.81 3.22
Unvegetated 0.52 0.12 0.01 0.1 Travel time [1-3[ 26.22 20.31 18.44 19.86
Savannas 94.9 98.56 80.04 85.51 (h) [3-10[ 53.77 59.16 65.26 62.55
[10-24] 15 17.17 13.48 14.37
[0-10[ 0.62 0.89 2.47 1.92
[0-1[ 0.03 0.02 0.48 0.33
Tree cover [10-40[ 62.04 81.57 90.02 84.59
Human [3-6[ 25.93 28.07 60.87 49.98
(%) [40-70[ 34.11 16.75 7.43 12.84
[70-100] 3.24 0.79 0.08 0.65
footprint [6-12[ 58 58.06 33.08 41.32 11
[12-50[ 16.03 12.84 5.58 8.37
 INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Local perceptions and adaptation strategies

66% 47 % 16%
1 3 5
Hunting Firebreak establishment Cooking at the field

50% 23% 6 10%

2 4
Agricultural plots establishment Charcoal production Unextinguished cigarette butts throw

Fig. 7: Causes of wildfires in Central Cote d’Ivoire 12

Kouassi et al. (2020)
INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Local perceptions and adaptation strategies

75 % 8%
1 3
Firebreaks establishment around new clearings Firebreaks construction around forests and fallows

10 meters
67 % 10%
2 4
Firebreaks construction and maintenance around farms Prohibition of fire-based hunting in the dry season
13
Fig. 8: Wildfires prevention strategies developed in Central Cote d’Ivoire Kouassi et al. (2020)
INTRODUCTION METHODOLOGY RESULTS PERSPECTIVES

  Challenges and future

Insufficiency of weather stations and climate data (2 synopotic stations in the Sudanian
zone with unavailable data during Ivorian political crisis period)
Fire risk index adapted to the country
Impact of wildfires on GHG emissions and fire emission baseline and estimate

14
African
Global
Impact
Raiser

Thank you for your kind

attention

37