SMOS L-VOD shows that post-fire recovery of dense forests is slower than what is depicted with X- and C-VOD and optical indices

Anthropogenic climate change is now considered to be one of the main factors causing an 10 increase in both frequency and severity of wildfires. These fires are prone to release substantial quantities of CO2 in the atmosphere and to destroy natural ecosystems while reducing biodiversity. Depending on the ecosystem and climate regime, fires have distinct triggering factors and impacts. To better analyse and describe fire impact on different biomes, we investigated pre and post fire vegetation anomalies at global scale. The study was performed using several remotely sensed quantities ranging from optical vegetation 15 indices (the enhanced vegetation index (EVI)) to vegetation opacities obtained at several microwave wavelengths (X-band, C-band, and L-band vegetation optical depth (X-VOD, C-VOD, and L-VOD)), ranging from 2 to 20 cm. It was found that Cand X-VOD are mostly sensitive to fire over low vegetation areas (grass and small bushes) or over tree leaves; while L-VOD depicts better the fire impact on tree trunks and branches. As a consequence, L-VOD is probably a better way of assessing fire impact on biomass. The study shows that L20 VOD can be used to monitor fire affected areas as well as post-fire recovery, especially over densely vegetated areas.

forests (de Faria et al., 2017). The Amazon could move from a net carbon sink to a net carbon source (Brando et al., 2020). These fire-prone conditions are largely attributable to anthropogenic climate change, and to human pressure (Weber and Stocks 1998;Randerson et al., 2006;Jin et al., 2012;Masrur et al., 2018; 65 Goldammer and Crutzen, 1993).
Naturally, understanding the potential factors triggering large uncontrollable fires is important to anticipate them and thus to reduce their impacts. Mhawej et al. (2015) presented and categorized 28 wildfire likelihood factors into climatic (e.g. precipitation, temperature, air humidity, wind speed), topographic (e.g. slope, altitude), in situ (e.g. fuel type, soil texture, tree diameter), historical, and anthropogenic factors. Drought, i.e. 70 the concomitant increase of air dryness and decrease of fuel moisture, was identified as the most significant fire likelihood factor (Ray et al., 2005). Indirectly, drought also reduces soil moisture, leading to leaf shedding and branch losses (Pausas and Bradstock, 2007). This process increases fuel accumulation and direct sunlight reaching the forest floor, increasing forest flammability (Nepstad et al., 2001). Surveying the soil moisture and the biomass status could then be a good indicator for fire detection. Currently, drought is mainly monitored 75 with temperature and precipitation observations. Fuel availability is monitored with optical sensors, but with fast saturation over dense vegetation, and lack of sensitivity for dry biomass (not green), which is the main fuel. Chu and Guo (2013) have reviewed the use of remote sensing in addition to field campaigns to evaluate the impacts of fire and post-fire vegetation regeneration. Indicators and metrics based on multispectral satellite 80 imagery (visible and infrared) are the most frequently used, such as the normalized difference vegetation index (NDVI), the enhanced vegetation index (EVI), and the normalized burned ratio (NBR) (Pérez-Cabello et al., 2021). Despite their fast saturation over dense forests, they still provide a good proxy for green vegetation regrowth. Microwave data also shown a good potential to monitor post-fire biomass recovery. Mermoz and Le Toan (2016) used L-band ALOS PALSAR to assess forest regrowth in South-East Asia. Fernandez-Carrillo et al.

85
(2019) used active L-band SAR data to estimate the tree survival in eucalyptus forests of Western Australia. Zhang et al. (2021) analysed forest canopy dynamics in the southern Amazon during the 2019 fire season using passive microwave vegetation optical depth at C-band (C-VOD). Authors found a larger magnitude and a faster recovery of optical-based indices (NDVI, EVI, NBR) than C-VOD.
With the arrival of L-band radiometers such as the Soil Moisture and Ocean Salinity (SMOS) satellite, it is 90 possible to infer complementary information, typically surface soil moisture, biomass (i.e. fuel) and its water content. The rationale of this study is to investigate how L-band radiometry responds to fire events over different ecosystems and under several climate circumstances. SMOS satellite has been operating for almost 12 years now and we have access to a large catalogue of major fires. This study also presents for the first time L-band used in conjunction with other sensors, from optical (EVI) to X-and C-bands. All the vegetation variables (VVs) derived from various sensors then represent different aspects of vegetation structure: canopy greenness for EVI, top canopy biomass for X-VOD (leaves and small branches), deeper canopy biomass for C-VOD (larger branches), and the whole AGB layer for L-VOD (stems and trunks) (Guglielmetti et al., 2007;Frappart et al., 2020). The complementarity of the derived VVs along with climate variables (CVs) (air temperature (T), precipitation (P), soil moisture (SM), and terrestrial water storage (TWS)) was used to identify 100 the fire likelihood factors, and the immediate and long-term fire impacts on vegetation. To do this, we first analysed three particular cases of large fires in various environments and then extended the analysis to the global scale.

105
Fires were obtained from the National Aeronautics and Space Administration (NASA) MODerate resolution Imaging Spectroradiometer (MODIS) Active Fire product (MOD14A1_M). The product is a quantification of the number of fires observed within a 1000 km² area over a month. A fire must cover at least~1000 m² to be detected. The Active Fire product is based on the 1 km fire channels at 3.9 and 11 μm of MODIS Terra and Aqua satellites (Justice et al., 2006). It is distributed at 0.1 deg resolution and at a monthly time scale by NASA 110 Earth Observations (NEO) portal.

Precipitation
Precipitation (P) data come from the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR). The precipitation estimate uses the PERSIANN algorithm on GridSat-B1 infrared satellite data, and training of the artificial neural network on the National 115 Centers for Environmental Prediction (NCEP) hourly precipitation data (Ashouri et al., 2015). The dataset is distributed by National Oceanic and Atmospheric Administration (NOAA) at a daily time scale, and at 0.25 deg resolution in the latitude band 60°S -60°N.

Soil Moisture
The soil moisture (SM) dataset comes from SMOS satellite, launched by the European Space Agency (ESA) in 120 2009 (Kerr et al., 2001). It performs passive measurements of the thermal emission of the Earth at L-band (1.4 GHz, 21 cm). L-band VOD and SM are derived from SMOS brightness temperatures using the L-band Microwave Emission of the Biosphere (L-MEB) radiative transfer model (Wigneron et al., 2007;Kerr et al., 2012). L-band SM is the volume of water per volume of soil (m 3 m -3 ) in the top surface soil layer (~5 cm). The https://doi.org/10.5194/bg-2021-289 Preprint. Discussion started: 11 November 2021 c Author(s) 2021. CC BY 4.0 License. footprint size is~43 km in average (Kerr et al., 2010). We considered the ESA level 2 SM dataset in version 125 7.2 (L2 v720) resampled to the global cylindrical Equal-Area Scalable Earth (EASE) Grid version 2.0 (Brodzik et al., 2012) at 625 km² spatial sampling (25 km × 25 km at 30 deg of latitude). Ascending (6 am) and descending (6 pm) overpasses were averaged, from June 2010 to December 2020.

Terrestrial Water Storage
Terrestrial water storage (TWS) anomalies from the Gravity Recovery and Climate Experiment (GRACE) 130 satellite were also considered. We used monthly GRACE/GRACE-Follow On (FO) Level-3 product provided through the Gravity Information Service (GravIS) web portal of the German Research Centre for Geosciences (GFZ) at 1 deg latitude-longitude grids (Boergens et al., 2019). TWS anomalies represent the water mass anomalies from snow, surface water, soil moisture, and deep groundwater. They are derived from measurements of temporal changes in the Earth's gravity field. Data were lacking for 35 dates of the ten-year 135 dataset. One-time gaps were filled by linear interpolation; consecutive missing months were not considered.

Temperature
Temperature (T) data come from the Land Surface Temperature (LST) dataset from MODIS Terra satellite (NASA). Daytime and night time measurements were averaged (MOD11C3 Version 6 product in a Climate Modeling Grid (CMG), LST_Day_CMG and LST_Night_CMG, Wan et al., 2015). These datasets are obtained using 140 MODIS thermal infrared bands from 3 to 15 μm, and distributed by NASA Land Processes Distributed Active Archive Center (LP DAAC) at a monthly time scale and at 0.05 deg resolution.

Vegetation Optical Depth
Vegetation optical depth (VOD) is a remotely sensed indicator related to AGB and to vegetation water content (VWC) (Kerr and Njoku, 1990;Jackson and Schmugge, 1991;Jones et al., 2011;Rahmoune et al., 2014;Vittucci 145 et al., 2016;Rodriguez-Fernandez et al., 2018;Mialon et al., 2020;Konings et al., 2021). The lower frequencies have better capabilities to penetrate deeper within the canopy (Ulaby et al., 1981). At L-band, VOD is sensitive to coarse woody elements, such as trunks, stems, and branches. At C-and X-band, VOD is more sensitive to thin stems and leaves (Guglielmetti et al., 2007). In this paper, L-VOD comes from SMOS level 2 dataset in version 7.2 (L2 v720) measured at 1.4 GHz (λ = 21 cm), resampled to EASE-Grid 2.0 at 625 km² resolution (25 150 km × 25 km at 30 deg of latitude). In the SMOS retrieval algorithm, the vegetation attenuation is taken into account by the τ parameter of the τ − ω model (Mo et al., 1982) which corresponds to the L-VOD. Data from June 2010 to December 2020 were considered, and ascending (6 am) and descending (6 pm) overpasses were averaged. C-and X-VOD from the Japan Aerospace Exploration Agency (JAXA) Global Change Observation https://doi.org/10.5194/bg-2021-289 Preprint. Discussion started: 11 November 2021 c Author(s) 2021. CC BY 4.0 License.

Year of gross forest cover loss event
The year of gross forest cover loss event map (the so-called lossyear product) from Hansen et al. (2013) was 170 used to observe the forest loss rate and year within a SMOS pixel, for the three major fires studied (Fig. 2).
This map represents the first year of detected tree loss during the period 2000-2020, defined as a standreplacement disturbance, or a change from a forest to non-forest state. This dataset is based on Landsat images and is distributed at~30 m resolution with 10x10 square degree tiles at https://glad.earthengine.app/view/global-forest-change. Each year of the period 2010−2020 was extracted 175 from the forest loss product and averaged into SMOS EASE-Grid 2.0, so as to obtain a yearly percentage of forest loss.

Above-Ground Biomass
The global map of AGB (Mg ha -1 ) from Santoro et al. (2021) was used to distinguish sparse from dense forests (see Sect. 3.2). This map is distributed through the ESA Climate Change Initiative (CCI) Biomass at 100 m 185 resolution. It combines a large pool of spaceborne remote sensing observations from two synthetic aperture radar (SAR) missions (Envisat and ALOS), and uses optical (Landsat) and LiDAR (Icesat GLAS) data to support the model calibration procedure. The ESA CCI Biomass map representative of 2010 was used here because it provides an AGB information prior to the studied fire events (2011-2020).

190
The Interactive Multisensor Snow and Ice Mapping System (IMS) database was used to mask areas covered with snow or ice (see Sect. 3.2). We used the IMS Daily Northern Hemisphere Snow and Ice Analysis at 4 km resolution, version 1 (Helfrich et al., 2007), provided by the National Snow and Ice Data Center (NSIDC).

Methods
First, we investigated three regional tests cases, diversely affected by human activities, from dry to wet climate, which recently suffered from a severe fire. These areas consist in: i) a eucalyptus open forest in a human-affected environment, under dry El Niño conditions in Australia; ii) a Mediterranean climate with needleleaf forests, agriculture, and human activities in California; and iii) a very dense primary forest in a 205 tropical wet climate with no human activities in Amazonia (see Sect. 3.1). Following this first step, the study was extended to the global scale, for five vegetation types, by selecting the major fires of the last decade. The rationale was to capture significant events occurring over an area large enough to be observed with the SMOS satellite without any ambiguity.
Four CVs related to the fire risk were considered: precipitation, SM, TWS, and temperature. Wind is another 210 predominant fire likelihood factor (Albini, 1993), but was not studied here due to the lack of global reliable https://doi.org/10.5194/bg-2021-289 Preprint. Discussion started: 11 November 2021 c Author(s) 2021. CC BY 4.0 License. database at a monthly time-scale. Vegetation status before, during, and after fire was monitored with four optical and microwave variables: EVI, X-VOD, C-VOD, and L-VOD. Data from June 2010 to December 2020 were considered (10.5 years), except for C-and X-VOD from AMSR-2 which were only available from July 2012.
Monthly averages of all datasets were computed and resampled to SMOS grid. To do so, they were first 215 oversampled to 1 km resolution and then averaged to SMOS EASE-Grid 2.0 (~25 km resolution) with GDAL (GDAL/OGR contributors, 2020).

Wildfires in the South Coast of New South Wales in Australia
The first studied area is located in the South Coast of New South Wales in Australia, between [33.53°S -  (McColl, 1969;DEWR 2007). The climate is warm temperate with dry summer (Kottek et al., 2006). The mean rainfall is~1000 mm year -1 , and the mean temperature is~15°C (McColl, 1969). The topography varies between 0 to 600 m above sea level. The 2019-2020 wildfires in Australia were influenced by El Niño Southern Oscillation (Dowdy, 2018). They became historically significant as they were widespread 225 and extremely severe, in particular in New South Wales (Ehsani et al., 2020). The tree cover loss map (Hansen et al., 2013) indicates a 25% forest loss in 2020 in the studied area (Fig. 2).

Mendocino Complex fire in California
The second studied area is located in California, between [38.96°N -39.46°N] and [122.68°W -123.20°W] (Fig.   1). It corresponds to four SMOS pixels. The area is covered with evergreen needleleaf forest and woody 230 savannas (Broxton et al., 2014), and is much urbanised. The climate is warm temperate (Kottek et al., 2006), with dry, windy, and often hot weather conditions from spring through late autumn that can produce moderate to severe wildfires (Crockett and Westerling, 2018). The 2018 fire season was the most extreme on record in Northern California (now second to the 2020 fire season) in terms of number of fatalities, destroyed structures, and burned areas (Brown et al., 2020). The largest fire complex in state history, the Mendocino complex, 235 burned nearly 186000 ha of vegetation between July and September 2018. The Mendocino complex started on July 27 near Lakeport lake and included two wildfires: the Ranch fire at the North of the lake, which was the largest single fire in state history, and the River fire at the West of the lake. The River Fire was fully contained on August 13 after having burnt 198 km 2 ; and the Ranch fire was contained on September 18 after burning 1660 km 2 (estimation from Bureau of Land Management). This wildfire was responsible for a 34% forest loss in 240 this region (26% in 2018 and 8% in 2019, Fig. 2), and caused a burn severity ranging from moderate to high. https://doi.org/10.5194/bg-2021-289 Preprint. Discussion started: 11 November 2021 c Author(s) 2021. CC BY 4.0 License.

Santarem wildfire in the Amazon
The third studied area is located in the Amazon rainforest near Santarem city (Brazil)

255
EVI, X-, C-, L-VOD, P, SM, TWS, and T anomalies time series were plotted over the three studied sites. The anomaly time series of a variable x is the difference between the original time series and the mean climatology, which is the monthly long-term mean of this variable. They are defined as: where t is the month number from January 2010 (6 to 132 in this study); m is the month of the year, between 1 and 12, with m = t mod 12 if t mod 12 > 0, and m = 12 if t mod 12 = 0; y is the year number, from 1 to yn, with yn = 11 here as the climatology was computed on the period 2010-2020. Plotting the anomalies time series enables to remove the natural seasonal cycle so as to observe only the variations due to specific events.

265
The average pre-fire variable value was subtracted from the anomalies time series, only if at least twelve months of data were available before the fire event. It enables to observe the anomalies with respect to the pre-disturbance value.

Extension to the global scale
Fires were then studied at the global scale to assess the general factors related to fire risk and the response of CVs pre-fire anomalies and VVs post-fire anomalies were also aggregated at different time frames and plotted by variable, in order to compare their pre and post fire temporal behaviour in different ecosystems.
The standard error of the mean of the measurements σ was also computed with: where std is the standard deviation of the population p and n is the number of samples. In evergreen forests of the South Coast of the New South Wales region in Australia (Fig. 3a), fires are 300 associated with high temperature and low precipitation (anom(T) = +3°C, anom(P) = -80 mm). The drought started 3 years before fire (decrease in precipitation, SM, and TWS). All VVs exhibit the same pattern, which is i) a constant and mild decrease since 2012, linked with the afore mentioned drought; ii) a strong decrease during the fire event (anom(VV)~-0.15); and iii) a rapid post-fire recovery (~1 year). C-VOD is the most affected VV.
The dense and humid rainforest near Santarem (Brazilian Amazon) shows stable VVs values before the December 2015 fire (Fig. 3c). L-VOD signal is quite noisy because only two SMOS pixels were considered here.

Extension to the global scale
In this section, the major fires from July 2012 to December 2020 were analysed at the global scale, by shifting the anomalies time series of all variables on the fire date tfire. The considered fires are well spread spatially and temporally over the nine-year period (Fig. 4). In savannas and grasslands ( Fig. 5a and 5b), pre-fire hydrologic variables are stable (positive for grasslands) and temperature anomalies are negative during 2 320 years before fire. Concurrently, VVs start to increase and reach a maximum a few months before the fire event (particularly C-and X-VOD over grasslands). The vegetation material surplus implies an increase of the available fuel, which may facilitate the propagation of wildfires. VVs anomalies also show a light surplus over needleleaf forests just before the fire event (Fig. 5c). Over broadleaf forests ( Fig. 5d and 5e), a pre-fire drought is visible through the temperature increase and the precipitation, SM, and TWS decrease, 12 months pre-fire 325 for sparse forests, 8 months pre-fire for dense forests. For all ecosystems, these drought conditions intensify just before and during fire, and end a few months after fire. During fire, all VVs abruptly decrease in all ecosystems, EVI being the most impacted one, excepted over dense forests where L-VOD heavily decreases (Fig. 5e). For all ecosystems, EVI recovers more rapidly than VODs. L-VOD is particularly long to recover over dense broadleaf forests (more than 4 years, Fig. 5e). Needleleaf forests (Fig. 5c)  fire in this ecosystem. In low vegetation ecosystems ( Fig. 5a and 5b), C-and X-VOD never regain their immediately pre-fire values, which were particularly high.
CVs anomalies were also averaged in space and in time, within time frames of 6 months, from 24 to 1 month pre-fire, in order to observe their general trends (Fig. 6). The error bars were computed with Eq. (3).

335
Precipitation anomalies (Fig. 6a) are negative from 6 months pre-fire for all classes, and reach -10 mm month -1 in average before the fire event. The precipitation deficit is more intense in dense broadleaf forests, where it starts 2 years pre-fire and reaches -65 mm month -1 before the fire event. SM anomalies (Fig. 6b) are similar for the three forest ecosystems. The SM deficit starts one year pre-fire and reaches -0.04 m 3 m -3 before the fire event. Savannas and grasslands are affected later (6 months pre-fire) and to a lesser extent (-0.01 m 3 m -3 ), as 340 previously observed in Fig. 5. TWS anomalies are negative from 12 months pre-fire, excepted for grasslands.
Again, dense broadleaf forests are the most impacted ecosystem, with a minimum TWS anomaly of -7 before fire. Temperature anomalies are not significant before 6 months pre-fire, excepted in grasslands and savannas where negative anomalies are visible. From 6 months pre-fire, temperature anomalies show a surplus in all ecosystems but savannas, and reach +0.9°C in needleleaf forests, +0.8°C in dense broadleaf forests, and 345 +0.5°C in grasslands before the fire event. In summary, pre-fire drought is mainly observed in forests, with particularly low hydrological values in dense forests (rainforests), and particularly high temperatures in needleleaf forests (boreal ecosystems). Savannas and grasslands barely suffer from pre-fire drought.
Temperatures are even mild one year pre-fire.
VVs anomalies were averaged within time frames of 6 months, from 1 to 36 months post-fire, in order to 350 observe the global impacts and recovery time (Fig. 7). We considered that a variable has totally recovered when its anomaly is between -0.005 and +0.005. Immediately after fire, EVI is the most impacted variable, with average anomalies of -0.06 over broadleaf forests, -0.04 over needleleaf forests and savannas, and -0.03 over grasslands (Fig. 7a). EVI recovers rapidly, in about 25 to 30 months over needleleaf forests, and 19 to 24 months over other ecosystems. X-VOD is less affected over forests (-0.02) than over low vegetation (-0.03) (Fig.   355   7b). X-VOD gets back to normal within 3 years, needleleaf forests and savannas being the longest to recover.
C-band is mainly impacted over forests and savannas, but barely over grasslands (Fig. 7c). C-VOD recovers slower than X-VOD, in particular over forests. L-VOD is mainly affected over broadleaf forests, and particularly over the densest ones (Fig. 7d). There, negative anomalies decrease up to -0.06 one year post-fire, then slowly increase. L-VOD is less affected than C-VOD elsewhere. It also shows a delayed impact by one year over 360 needleleaf forests, as for C-and X-VOD.

380
In Australia ( fig. 3a), Ehsani et al. (2020) explained that the air temperature from December 2019 to February 2020 was about 1°C higher than usual, which increased evapotranspiration. The pre-fire lack of precipitation, due to a high-pressure system over South-East Australia, prevented the soil from supplying the moisture demand, and led to a significant vegetation drying (fuel) that facilitated the propagation of fires. L-VOD regained its pre-fire values within a year, meaning that the woody biomass was not entirely destroyed. Indeed, 385 these eucalyptus forests are known to be somewhat fire resistant (Wilkinson and Jennings 1993;Caccamo et al., 2015). They can regenerate branches and leaves by resprouting from heat-resistant buds (Burrows, 2002).
The rapid recovery of VVs could also be explained by an increase in vegetation water content, linked with the post-fire increase in precipitation and SM (Konings et al., 2021). In 2020, SM values exceeded those of the previous decade (anom(SM) = +0.15 m 3 m -3 ), corresponding to the end of the severe drought affecting South-

390
East Australia associated with the 2020/2021 La Niña event (Australian Bureau of Meteorology). The increase of precipitation may also have expedited the extinction of fires (Ehsani et al., 2020).
In California, all climate and fuel conditions leading to the 2018 widespread Mendocino complex described in Brown et al. (2020) are visible in Fig. 3b. We can observe positive rainfall and SM anomalies in winter 2016/2017, which led to the second consecutive spring with above average accumulation of fine fuel (grasses).

395
Then, we can observe positive temperature anomalies in winter 2017/2018, when a lack of storm enabled the survival of grasses. In April 2018, precipitation and warm temperatures led to above normal spring brush and grass growth. No major drought is visible in summer 2018, but low rainfall and warm temperatures led to a rapid drying of fuels, and induced a poor overnight humidity recovery. The dramatic fire impacted EVI and L-VOD in the long term. Eucalyptus, pine trees and chapparal were burnt. Even if this type of vegetation is fire-400 adapted, the severity of the fire seemed to have destroyed most of it. Increased forest fire activity in recent decades has likely been enabled by the legacy of fire suppression, human settlement, and anthropogenic climate change (Abatzoglou and Williams, 2016). Stephens et al. (2018) stated that the massive current tree mortality in California will undoubtedly provoke severe "mass fires" in the coming decades, driven by the amount of dry and combustible wood.

405
The wildfire in Santarem region (Fig. 3c) was attributed to high temperature and low precipitation linked with El Niño event in winter 2015 (Berenguer et al., 2018). These extreme drought conditions worsened during the fire, and may explain its severity. The 3-year recovery time of EVI after the severe fire indicates a moderate regrowth of leaves and grasses. In contrast, L-VOD never regained its pre-fire values, meaning that trunks were impacted on the long term.

Extension to the global scale
At global scale, grasslands, tropical and subtropical savannas do not show signs of pre-fire drought (Fig. 5a, 5b,   6). A substantial increase in VVs occurs 1-2 years before fire, which implies an increase in the available fuel, and may contribute to trigger the wildfires. Indeed, the fire risk in savannas is highest for a dry vegetation with enough fuel to enable a drastic burning (Mbow et al., 2004). This vegetation growth might be enabled by 415 negative pre-fire temperature anomalies, and light positive pre-fire hydrological variables anomalies (Fig. 6).
VVs are less impacted by fires (Fig. 7) because they are rapid and burn through the grass layer, resulting in less destruction than in forests (Menaut, 1983;Bessie and Johnson, 1995;Gignoux et al., 1997). L-VOD in particular is slightly impacted because the burnt vegetation is mainly leaf biomass. EVI quickly recovers after fire, probably because fire burns most of the AGB of grass species, but spares their large underground root 420 systems, resulting in a rapid establishment of new shoots (Hochberg et al., 1994). The extremely high pre-fire VVs values are never regained.
In needleleaf forests time series (Fig. 5c), the numerous missing values correspond to the filtering of snow in winter. Wildfires in this biome are located in the Northern hemisphere temperate/boreal forests (Fig. 4a), and mostly occur in late spring and in summer (Fig. 4b).
De Groot et al. (2013) stated that most boreal fires in

425
Canada are due to lightning strikes and occur during summer; whereas most fires in Russia are human-caused and occur during spring. We found a strong pre-fire drought in this ecosystem (low SM, high T, Fig. 6), which is well documented (Weber and Stocks, 1998). VODs also showed a light surplus before fire, possibly linked with litter thickening, which also increases the fire risk. Indeed, de Groot et al. (2013) explained that in Russian boreal forests, extensive areas of dead light surface fuels (e.g. dead needles, cured grass, leaf litter) are 430 available in spring and facilitate fire propagation. We found a delayed impact of fire on VVs, and a longer recovery time than in other ecosystems, of about 3-4 years ( Fig. 5c and 7). Based on NDVI, Goetz et al. (2006) found a 5 year recovery time in Canada. Jin et al. (2012) found that the summer EVI has recovered in 5 to 8 years after fire. These findings also confirm previous results from Yang et al. (2017), who showed with NDVI analyses over North America that the fire effect on high latitude needleleaf trees was stronger and longer than 435 on other vegetation types. Indeed, in North America, boreal forest fires are predominantly stand-replacing and high-intensity crown fires (Stocks et al., 2004;Jin et al., 2012). In Eurasia, most fires occur as surface fire and are usually of low to moderate intensity (de Groot et al., 2013). We found that L-VOD is less impacted by fire than other VVs in this biome. It can be explained by a dominant destruction of canopy leaves and branches by crown fires (Alexander and Cruz, 2011).

440
Sparse broadleaf forests (AGB ≤ 150 Mg ha -1 ) subject to wildfires are mostly located in subtropical and temperate areas of South America, North America, West Africa, and South-East Asia (Fig. 4a). A drying trend is visible 1 year pre-fire ( Fig. 5d and 6). The link between drought and wildfires was previously observed by de  Marzo et al. (2021) in the Argentine Gran Chaco; by Cheng et al. (2013) in the Mexican Yucatan forest; and by Vadrevu et al. (2019) in South-East Asia, with a prominent influence of precipitation variations over 445 temperature variations. L-VOD and EVI are particularly impacted by fire, but they recover quickly (1 year for EVI, 2 years for L-VOD). Yang et al. (2017) also found a rapid recovery time over North American broadleaf trees due to their fire-adaptive resprouting regeneration mode. Same observations were made in the fireprone Argentine Chaco forest by Torres et al. (2014).
Dense broadleaf forests are mostly located in the tropics (Fig. 4a). We can notice few fires in the densest 450 rainforests (Congo basin, central Amazon) as they are usually too humid to burn (Cochrane, 2003); and because seasonally flooded areas were excluded. A consistent drought is visible 8 months before fire (Fig. 5e).
We observed high negative SM, TWS, and precipitation anomalies in particular (Fig. 6). Chen et al. (2013) also found TWS deficits before severe fire seasons across the southern Amazon. Drought-related fires were even suggested to prevail over deforestation fires in the Amazon, and are predicted to increase in the near future 455 (Aragão et al., 2018). Indeed, rainfall shortage generates high water deficits (i.e. high negative SM anomalies), which causes tree mortality, leaf shedding (visible in pre-fire EVI decrease) and thus increases fuel availability.
The opening of forest canopies also boosts incident radiation levels which leads to temperature rise (Ray et al., 2005). The combination of fuel increase in a drier and hotter environment converts natural forests into fireprone ecosystems (Aragão et al., 2018). We also found that the dense broadleaf forest biome was the most 460 impacted by fire ( Fig. 7), because it is not a fire-adapted ecosystem (Cochrane, 2003). L-VOD in particular decreases strongly and recovers very slowly (Fig. 7d), as previously observed over Santarem fire (Fig. 3c). The strong post-fire decrease in L-VOD is due to biomass destruction but also to water stress in the remaining vegetation (Konings et al., 2021). This finding confirms the significant and damaging impact of fires in the dense broadleaf ecosystem previously observed by Silva et al. (2018) andde Faria et al. (2021). L-VOD was 465 previously proven to be more sensitive to high AGB values than C-and X-VOD (Rodriguez-Fernandez et al., 2018). Here, we suggest that L-VOD depicts better the fire impact on high AGB areas than the other VVs.
For all biomes, EVI is the most rapid index to recover, because leaves rapidly resprout. EVI and X-VOD seem better adapted for grasslands fire monitoring, C-VOD for savanna fire monitoring, and L-VOD for forest fire monitoring.

The potential of L-VOD for vegetation recovery monitoring over dense forests
VV anomalies were also plotted with respect to the number of fires in the dense broadleaf ecosystem, immediately after fire (from 1 to 3 months post-fire, Fig. 8a) and over a longer period (from 1 to 2 years postfire, Fig. 8b). A quasi-linear relationship is visible for all VVs. As previously observed in Sect. 4.2, EVI and L-VOD are the most impacted variables immediately after fire (Fig. 8a), while L-VOD is still significantly affected 1 to 2 years after fire (Fig. 8b). L-VOD then shows a clear response to fire events over high AGB areas, immediately after fire but also in the long term, and proportionally to the number of fires within a SMOS pixel. Thanks to its sensitivity to coarse woody elements and to its deep penetration through the vegetation layer, L-VOD is better correlated to high AGB than other VVs (Rodriguez-Fernandez et al., 2018), and could be used for post-fire recovery monitoring over dense forests.

Conclusion
In this paper, we analysed the pre-fire behaviour of four fire likelihood factors, including SMOS SM which provides access to top surface soil moisture at the global scale. In forests, which generally maintain a steady humidity, we found that fires are linked with intense and prolonged drought. Pre-fire temperature anomalies are particularly high in boreal needleleaf forests. In savannas and grasslands, in agreement with previous 490 studies ( Mbow et al., 2004), we found evidences of an increase in available fuel prior to fire events, enabled by humid and cold conditions a few years before. We also found that vegetation variables recover rapidly in these ecosystems, as wildfires are often rapid and mildly destructive for trees. In contrast, over forests, fires can damage the vegetation in the long term. Zhang et al. (2021) demonstrated the potential of C-band vegetation optical depth to detect the vegetation change patterns caused by fire in the southern Amazon. Our study 495 confirms these findings and extends it to the global scale, and to two extra wavelengths. Dense broadleaf forest fires particularly impact the L-band emission, which represents coarse woody elements (trunks and stems); whereas sparse vegetation fires affect more C-and X-bands, which are more sensitive to small branches and leaves. For all biomes, the optical-based index (EVI) drops down after fire but recovers quickly,

Competing interests
The authors declare that they have no conflict of interest.