What is the potential of cropland albedo management in the fight against global warming? A case study based on the use of cover crops

Dominique Carrer; Gaétan Pique; Morgan Ferlicoq; Xavier Ceamanos; Eric Ceschia

doi:10.1088/1748-9326/aab650

1. Introduction

Today, the scientific community has evidence that the global warming issue (IPCC 2014) will not be resolved without clever solutions to reduce human-induced radiative forcing (RF). Possible strategies for climate mitigation include carbon dioxide removal (CDR) approaches or solar radiation management (SRM). While CDR aims at removing some CO₂ from the atmosphere, the goal of the latter approach is to counteract greenhouse gas-induced warming with an increase in the sunlight reflected back to space by an increased average albedo of the Earth. Lenton and Vaughan (2009) estimated that SRM strategies have the potential to cool the climate to its preindustrial state.

Table 1. Different possible types of crop rotations during a 3 year period: summer–summer–summer (Rsss), winter–winter–winter (Rwww), summer–winter–summer (Rsws), and winter–summer–winter (Rwsw). In this example, the summer and winter crop periods are represented by black and grey, respectively. The period and duration of the vegetation cycles are estimated according to the approximate seeding and harvest dates of each crop over central and western Europe. Green boxes show favourable periods for the introduction of cover crops that have been considered in this study. Note that this period and its duration vary spatially in Europe.

Various initiatives have been proposed to control the Earth's albedo. Although the dispersion of sulfate aerosols into the atmosphere is the most emblematic, this approach could have unintended and possibly harmful consequences (Robock et al 2009). Changing the surface albedo naturally seems to be a more appropriate solution. Some studies have suggested that an increase in surface albedo can be achieved by making roofs white worldwide (Akbari et al 2009, Jacobson and Hoeve 2011). Other initiatives support deforestation or forestation strategies to induce surface albedo changes (Betts 2000, Davin et al 2014, Singarayer and Davies-Barnard 2012). An advantage of modifying the continental surface albedo for climate mitigation purposes (Lenton and Vaughan 2009) is that its progressive implementation could help limit the risk of an excessively rapid or strong negative climatic response and could, therefore, be reversed.

This study analyses the radiative forcing changes induced by an increase in the surface albedo of croplands in Europe following the introduction of a cover crop during the fallow period, that is, between the harvest of winter crops and the seeding of the following summer crops. After the harvest, croplands are often bare soil, which has usually a lower albedo than vegetation (Aguiar and Page 1999, Campbell and Norman 1998, Davies and Idso 1979, Oke 1987, Carrer et al 2014). The seeding of a cover crop (also referred to as a catch crop or green manure crop) during the fallow period, in places where the period is long enough, would permit more solar energy to be returned to space than when the soil remains bare after the harvest and during the entire winter season (Kaye and Quemada 2017). Many types of plants can be used as catch (or cover) crops. The most extensively used are legume (fava bean, clover) and grasses, but there is increasing interest in brassicas (such as rape, mustard, and forage radish). Another potential advantage of this practice is that it may allow the significant storage of carbon in the form of organic matter in the ground (Justes et al 2012, Poeplau and Don 2015), thus combining SRM and CDR approaches (Smith and Rasch 2012), as recommended by the IPCC (2014).

The main objective of this article is to investigate if the inclusion of cover crops in crop rotation is beneficial for climate mitigation purposes and if, therefore, this procedure could be considered by the Europe Commission to counteract climate change in the 28 Member States (EU-28). To this end, we quantify first the potential of global warming compensation resulting from human-induced surface albedo increases due to the introduction of a cover crop at the European scale. In a second part of this work, we also discuss the impact of cover crops on soil carbon sequestration and the use of fertilizer. For this purpose, we use satellite data, land cover databases, meteorological data from model reanalyses, ground measurements, and national agricultural statistics. The potential of these cover crops to increase latent heat fluxes at the expense of sensible heat fluxes, and to decrease surface temperature is not discussed here, even if these processes may increase the mitigation effect (Kaye and Quemada 2017).

2. Methods and materials

2.1. Approach to cover crop inclusion

This section identifies the areas in Europe where cover crops could be sown during the fallow period. Only agricultural areas with annual cropping are included in this study (section 2.1.1). Furthermore, the favourable periods for the introduction of cover crops are not the same everywhere (section 2.1.2).

2.1.1. Location of potential areas for the inclusion of cover crops

Suitable areas are those containing croplands with annual cropping. This includes the two families of arable crops that are predominant in central and southern Europe, summer crops (sown in spring and harvested in autumn) and winter crops (sown in autumn and harvested in early summer). Table 1 presents the four main crop rotations that exist. As the table shows, the duration of the fallow period can either be too short (Rwww, winter–winter–winter case) or be associated with a late and unfavourable seeding period for cover crops (Rsss, summer-summer-summer case). Consequently, cover crops are mainly implanted between a winter crop (after) and a summer crop (before) in Europe (see green boxes, table 1); thus, Rsws (summer-winter-summer case) and Rwsw (winter-summer-winter case) were the only cases considered in this study to quantify the mitigation potential of cover crops.

To estimate the location of croplands in Europe from year to year at a fine spatial resolution, as well as their associated crop rotation type, we used ECOCLIMAP land cover (Masson et al 2003, Faroux et al 2013) and the European agricultural statistics (Eurostat) in 2011 (http://ec.europa.eu/eurostat/web/agriculture/data/database; last consulted: June 2016). ECOCLIMAP includes 520 ecosystems, or cover types, that are defined at a spatial resolution of 1 km. The heterogeneity of an ECOCLIMAP grid cell includes the mixing of 11 co-existing vegetation types, which may include winter and summer crops. Since ECOCLIMAP does not provide information about the possible crop rotations, the percentages of summer and winter crops and the crop rotation ratios (Rsws, Rwsw, Rsss, Rwww) in the sub-pixels were refined based on agronomical expertise and the 2011 national statistics from Eurostat. By doing this, the correct proportions of winter and summer crops were obtained for each country. Finally, we determined that the winter–summer crop rotation was the most common crop rotation in Europe (26% of the crop rotations).

2.1.2. Period in which a cover crop can be grown

This study analysed data from three years, from 2008–2010, during which the introduction of cover crops in the fallow period was considered. To identify the winter and summer crop harvest and sowing times for each grid cell, we used a method proposed by Gibelin et al (2006) and Szczypta et al (2012) based on the vegetation index obtained from the ECOCLIMAP database. First, the harvest was estimated to occur in the declining phase of the vegetation index, when it decreases below 40% of the yearly maximum. Second, seeding was estimated to occur when the vegetation index begins to increase the following year. We used the ECOCLIMAP vegetation cycle climatology derived from satellite observations (Faroux et al 2013) to estimate these two occurrence dates for the winter and summer crops in each grid cell. For areas where winter to summer crop rotation occurs (mostly in central and southern Europe), the estimated fallow period between the winter crop harvest and the summer crop seeding was considered to be the potential period for the cover crop introduction.

2.2. Radiative forcing

2.2.1. Relationship between surface albedo and TOA radiative forcing

In the remote sensing of continental surfaces, the total surface albedo of a given area (or a pixel, in our case) throughout a full year can be expressed daily as the weighted sum of the vegetation albedo (crop, in our case) and bare soil albedo:

$\begin{equation} \begin{array}{l@{}l@{}l@{}} \alpha (d) = (1 - {\rm veg}(d))\alpha_{bs} (d) + veg(d)\alpha_{\rm veg} (d)\\ \alpha_{\rm CI}(d) = (1 - {\rm veg}(d) - {\rm veg}_{\rm CI}(d))\alpha_{bs}(d)\\ + {\rm veg}(d)\alpha_{\rm veg} (d) + {\rm veg}_{\rm CI} (d)\alpha_{\rm vegCI} (d), \end{array} \end{equation} \tag{ 1 }$

where α, α_bs, and α_veg are the total, bare soil, and vegetation (summer or winter crop here) albedos, respectively. The parameter veg is the vegetation fraction. If a cover crop is added in the crop rotation, a vegetation fraction (veg_CI) of the cover crop is added to the equation (see the second line of equation 1), and the total albedo (α) becomes α_CI. α_vegCI is the vegetation albedo of the cover crop. Calculations for total albedo are done on a daily basis (d).

The direct radiative forcing in W.m⁻² at the top-of-atmosphere (TOA) level due to the change in the surface albedo (RF_Δα), here caused by the introduction of a cover crop in the crop rotation, is expressed in units of time for each pixel, as follows (Lenton and Vaughan 2009):

$\begin{equation} \begin{array}{l@{}l@{}l@{}} {\rm RF}_{\Delta \alpha } ({\rm W.m}^{ - 2} ) = - 1/{\rm Ndays}\sum\limits_{d = 1,{\rm Ndays}} {\rm SW_{in} } (d)\\ \times \, T_a (d) \times \Delta \alpha (d) \\ {\rm with}\;\Delta \alpha (d) = {\rm CI}\_{\rm Ratio} \times (\alpha _{{\rm CI}} (d) - \alpha (d)), \\ \end{array} \end{equation} \tag{ 2 }$

where SW_in is the total incoming solar radiation at the surface, T_a is the upward atmospheric transmittance, ∆α is the variation in surface albedo, and CI_Ratio is the percentage of cover crop introduction in the Rsws or Rwsw rotation. It is fixed between 0 and the maximum value of the crop rotation ratio (Rsws or Rwsw) in the sub-pixels (according the favourable periods and areas defined in sections 2.2.1 and 2.2.2). RF_Δα is the annual average of the daily radiative forcing (d from 1 to Ndays, where Ndays is equal to 1095 in this study, i.e. 3 entire years). The value of the RF_Δα at the TOA level is representative of the daily local power in W.m⁻² that would be reflected back to space due to the introduction of a cover crop. The estimation of RF_Δα resulted from 3 years of data, from 2008–2010, and was calculated based on the daily values of radiative forcing. As a matter of fact, all parameters (the albedo values, the vegetation fractions, the incoming solar radiation, etc.) in equation 2 were considered on the daily basis. This means the direct RF_Δα over 3 years was calculated for each pixel grid over Europe. The methods for obtaining all the parameters in equation 2 are described in the following section.

2.2.2. The α_bs, α_veg, and veg data

In the last decade, surface albedo estimates have become available at the global scale using satellite observations from different instruments (Qu et al 2015). These instruments allow us to estimate the Earth's surface albedo at a spatial resolution between 500 m and 5 km, with usually less than 10% uncertainty (Carrer et al 2010). Given the several types of land covers that co-exist in each satellite pixel at these resolutions, Carrer et al (2014) attempted to improve the characterization of the heterogeneity of the grid cells. This was done by developing a mathematical method based on ECOCLIMAP prior information (see section 2.1.1) to derive the surface albedos of up to 11 co-existing vegetation types (grassland, broadleaf, evergreen, summer crop, winter crop, etc.) and a bare soil in the same grid cell. In addition, the method proposed by Carrer et al (2014) allowed the effective capture of all seasonal or intra-annual and inter-annual albedo fluctuations of up to 12 pure, co-existing vegetation cover types and their underlying soils in the same pixel grid. The ECOCLIMAP land cover database was here used to determine the fraction of each co-existing vegetation type in a same grid cell. These fractions were readjusted at the country scale with the inventoried fractions of the different crop types provided by the 2011 national statistics from Eurostat (see section 2.1.1). In equation 1, veg, α_bs, and α_veg were obtained by summing the different contributions from the pure vegetation types (maximum of 12 co-existing types) that exist at the sub-pixel scale. Values corresponding to the pure vegetation characteristics at the sub-pixel scale were from Carrer et al (2014). Only the fraction of area covered by the summer or winter crop in a given grid cell was potentially impacted by the introduction of a cover crop in equation 1. In this grid cell, the temporal fluctuations of the bare soil albedo were distinguished from the albedo changes of the different vegetation types. Updates of these values were conducted with the MODIS satellite product (MCD43GF) of the snow-free albedos (bi-hemispherical albedos in the shortwave domain, [0.3–4 µm]) using the Kalman filter method. Again, Carrer et al (2014) provided, for the first time, estimates of the temporal evolution of bare soil albedo and vegetation albedos of crops at a global scale (which is necessary to properly conduct this study). These time series from 2008–2010 were used in the present study (albedo data of winter and summer crops and of bare soil - α_veg and α_bs).

2.2.3. The α_vegCI and veg_CI data

We interposed the development of some vegetation during the fallow period into the crop rotation between the winter and summer crops (see green boxes in table 1). The fallow period was determined according to the method presented above (see section 2.1). The maximum fractional presence of the cover crop (veg_CI) was arbitrarily fixed to 0.95×max(veg). It was assumed here that the level of development (or abundance) of the cover crop will not exceed the maximum level of development of the crop in a given location, a conservative approach. The vegetation fraction of the introduced cover crops gradually increased to the maximum value above. We used a linear interpolation to simulate this increase in the vegetation fraction, corresponding to the growing phase of the crop cover. Then, veg_CI remained constant until the cover crop was removed (veg_CI set to zero). Different scenarios for the date of removal were tested (see section 2.2.6). Equation 1 was used to estimate each daily value of the crop cover albedo (α_CI) derived from the daily estimates of veg_CI combined with α_bs, and α_vegCI. The value of α_vegCI was arbitrarily fixed at 0.95×max(α_veg), although it could be higher, according to Ferlicoq (2016).

2.2.4. The SW_in and T_a data

The incoming solar radiation at the top of the atmosphere (SW_TOA) and at the surface level (SW_in) reported by the ECMWF (European Center for Medium Range Weather Forecasts) reanalysis (Berrisford et al 2011a, Berrisford et al 2011b, Dee et al 2011) were used in this study. Assuming the upward and downward atmospheric transmittances to be equal, T_a in equation 2 was approximated as the ratio SW_in/SW_TOA. In comparison, when a mean annual and spatially constant upward T_a of approximately 0.85 is used as in Lenton and Vaughan 2009 and Kaye and Quemada 2017, that value has a tendency to overestimate the RF_Δα. In addition, the rainfall data from this reanalysis was used later in section 3.3 to consider the water needs of the cover crop emergence. Details concerning the satisfactory quality of these ECMWF fields are given in Dee et al (2011) and Szczypta et al (2012).

2.2.5. Conversion of radiative forcing into equivalent CO₂

Hereafter, we present two methods to convert radiative forcing into equivalent CO₂. The estimations delivered by the two methods will be compared in section 3.

Method 1 (based on a constant CO₂ airborne fraction, AF)—To compare the previously obtained RF_Δα with sources of CO₂ emissions, Betts (2000), Bird et al (2008), Munoz and Campra (2010) and Bright (2015a) convert RF in W.m⁻² into kgCO₂−eq.year⁻¹, as follows

$\begin{equation} \begin{array}{l@{}l@{}l@{}} {\rm RF}_{{\rm CO}_2} ({\rm kgCO}_2 - {\rm eq}.{\rm year}^{ - 1} ) = \\ \frac{{S \cdot {\rm RF}_{\Delta \alpha } }} {{\rm {AF}}} \times \frac{{\ln 2\rm {pCO}_{2,\rm {ref}} \cdot M_{\rm {CO}_2 } \cdot m_{\rm {air}} }} {{S_{\rm {Earth}} \Delta F_{2X} M_{\rm {air}} }}\frac{1} {{\rm {TH}}}, \\ \end{array} \end{equation} \tag{ 3 }$

where S is the area affected by the change in surface albedo (in m²), RF_Δα is the radiative forcing at the TOA level (in W.m⁻²; see equation 2), pCO_2,_ref is a reference partial CO₂ pressure in the atmosphere (383 ppmv), M_CO₂ is the molecular weight of CO₂ (44.01 g.mol⁻¹), m_air is 5.148 × 10¹⁵ Mg, S_Earth is the area of the Earth (5.1 × 10¹⁴ m²), ΔF_2X is the radiative forcing resulting from a doubling of current CO₂ concentration in the atmosphere (+3.7 W.m⁻²), and M_air is the molecular weight of dry air (28.95 g.mol⁻¹). With equation 3, the local power in W.m⁻² due to the albedo change (equation 2) that we estimated at the European spatial scale (surface area S) was converted into global RF_CO₂ (surface area S_Earth). TH is the time horizon. Based on the recommendations of Anderson-Teixeira et al (2012) and Kaye and Quemada (2017), the time horizon of our potential global warming calculations was fixed at 100 years (which supposes that cover crops will be maintained for this duration during the fallow periods). The per-year CO₂-eq from the albedo change was 1/100th of the total CO₂-eq due to the albedo change. In this way, the estimates of an equivalent CO₂ pulse due to the albedo change can potentially be compared to other sources of CO₂ emissions (for example, the energy, agriculture, or transport sectors). Note that the short analysis times of the cover crop introduction in the crop rotations overemphasize the albedo effect, while long analysis times, such as that in this study, deemphasize this effect (Anderson-Teixeira et al 2012). More studies are needed to determine the most appropriate time frame for this analysis; this is currently an active area of research in environmental biophysics (Bright et al 2015b).

Parameter AF is the average CO₂ airborne fraction, defined as the ratio of the annual increase in atmospheric CO₂ to the total CO₂ emissions from anthropogenic sources. In other words, it represents the proportion of human-emitted CO₂ that remains in the atmosphere after a certain period of time. Considering the Bern carbon cycle model (Joos et al 2001), after 10 years, 66% of the initial emission remains in the atmosphere due to CO₂ decay over time, while only 36% remains after 100 years. The integral of Bern carbon cycle model gives a 100 year AF value of 0.48 (quite close to 0.5 and 0.55 used by Betts (2000) and Akbari et al (2009), respectively). If all variables taking constant values in equation 3 (right-hand term) are grouped into a single parameter (rf_CO₂ = 0.908 W.kg.CO₂⁻¹), we obtain the following, according to Munoz et al (2010) and Bright et al (2015b):

$\begin{equation} \begin{array}{l@{}l@{}l@{}} \rm {RF}_{\rm{CO}_2 } (\rm{kgCO}_2 - \rm{eq}.\rm{year}^{ - 1} )\\ = \frac{{\rm{RF}(W.m^{ - 2} ) \times S(m^2 )}} {{\rm{AF} \times \rm{rf}_{\rm{CO}_2 } (W.kg_{\rm{CO}_2 }^{ - 1} )}}\frac{1} {{\rm TH}} \end{array}, \end{equation} \tag{ 4 }$

where AF is the atmospheric fraction for a 100 year time horizon (TH = 100) and rf_CO_₂ is the derived radiative forcing from 1 kg of CO₂. The uncertainty in these calculations depend on the respective uncertainties of α_bs, α_CI, SW_in, and T_a in equation 2, as well on rf_CO_₂ and AF when RF is converted to kgCO₂−eq. For rf_CO_₂, Akbari et al (2009) suggest a ±10% error whereas the error concerning AF is less than ±15% according to Forster et al (2007).

Method 2 (Global Warming Potential)—The use of a constant AF does not represent the variations in the emission rates of atmospheric CO₂, which are non-negligible over a 100 year period. The Global Warming Potential method (GWP method, IPCC's emission metrics, Myhre et al (2013)), which was also used in this study, attempts to take into account these variations in the atmospheric carbon concentration by using impulse-response functions (IRFs) (Joos et al (2013), Myhre et al (2013)). The converted CO₂-eq(t) decreases rapidly in the short term but very slowly over the long term. In the same way as above, to obtain a per-year CO₂-eq, we divided the 100 year GWP by 100. Still, as mentioned above, the short analysis times of the cover crop introduction in the crop rotations overemphasize the albedo effect, while long analysis times, like that used in this study, deemphasize this effect.

2.2.6. Scenarios for introducing the cover crops

In the first scenario that we tested, the cover crop was added during the first three months following the harvest of the winter crop, when possible. This three-month period was tested first, as it corresponds to the duration of cover crop introduction period that is recommended in some European countries to limit nitrate pollution when cover crops are used as catch crops. In the second scenario, we accounted for limitations due to the water requirements of cover crops. The rainfall values in each pixel of our study grid were used to limit the area where the crops could grow (figure 2(c)). Rainfall data from the ECMWF reanalysis were used with a threshold of 50 mm (the cumulative value for the first month after seeding) for the development of the cover crop. This condition is more restrictive than the requirement of 30 mm, which was estimated by Brisson et al (2009). All zones where rainfall is lower than 50 mm were excluded from our calculations in this second scenario. In the third scenario, we calculated the greatest impact by extending the cover crop for a period longer than 3 months, i.e. the longest possible period up to a maximum of 6 months, depending on the duration of the fallow period for each pixel. No limitation due to the water supply was introduced here. In all scenarios, the cover crop was not added if the introduction period was less than 1 month.

**Figure 1.** Mean bare soil albedo in August 2008 (a) and, December 2008 (b) and soil types from the HWSD database (c). The black rectangle delimitates the chernozems area (dark blue). A picture of this soil is shown in (d).
Download figure:
Standard image High-resolution image

**Figure 2.** (a) Starting date in days after the 1st of January and (b) duration in months of the 2008–2009 fallow period for the winter–summer crop rotation (between year 1 and year 2, figure 1 and table 1). (c) Mean cumulative precipitation during the month following the starting date of the fallow period (or the seeding of the cover crop), see (a).
Download figure:
Standard image High-resolution image

**Figure 3.** Mean daily value of (a) Ta ⁎ SW_in (W.m⁻²), (b) the increase in the surface albedo with the introduction of a cover crop, and (c) the radiative forcing (RF_Δα) due to the introduction of a cover crop for the 3 months after the winter crop harvest in 2008 (in W.m⁻²).
Download figure:
Standard image High-resolution image

3. Results

3.1. Albedo changes

The albedo of bare soil may change with time. In fact, values are usually lower in the winter than in summer due to an increase in the soil water content. Figure 1 shows a comparison between the bare soil albedo values retrieved in August and December 2008. These estimates were derived from MODIS satellite data (Carrer et al (2014)). The darkening of soils is one of the most important factors determining how profitable the introduction of the cover crop could be. The areas where the soil was brighter than a given threshold (typically 0.2) in figures 1(a) and (b) are likely unsuitable for introducing any cover crop. Indeed, the albedo of crops is typically between 0.15 and 0.3, except for some areas, such as Spain, that show a higher bare soil albedo in the summer. Figure 1(c) shows the different types of soil classified by the Harmonized World Soil Database (HWSD—Version 1.21; Fischer et al (2008)). The darkest soil in the world is the chernozem type (in dark blue, figure 1(c)). There are two 'chernozem belts' in the world, and the main one is located north of the Black Sea (see figure 1(b)). The albedo of chernozem soil becomes very low in this area in December, and during this period, the soil is often bare due to the fallow.

3.2. The fallow period

We estimated the start date of the fallow period for the winter-summer crop rotation following the methodology presented in section 2.1. This occurs across Europe, on average across, on day 241 (August 29). The fallow period occurs earlier in the southwest than in the northeast of Europe (see figure 2(a)). The latest fallow periods start in the northern latitude regions (i.e. Norway, Sweden, and Finland) and over mountainous areas (e.g. the Carpathian, Alpes, and Pyrenean Mountains).

Figure 2(b) shows the duration of the fallow period between the winter crop and the following summer crop. The average duration of the fallow period is 5.96 months, and it typically varies from 9 months in the southwest (in Spain) to 5 months in the northeast (in the Baltic States) of Europe. Hence, the duration is longer than the 3 months, which corresponds to the first and second tested scenarios (see section 2.2.6), nearly all over Europe, except for in Scandinavia, where it lasts approximately 2 months. It is important to remark that cover crops could be added for periods exceeding 3 months in multiple areas in Europe.

The starting date of the fallow period (figure 2(c)) can also correspond to the date of seeding of the cover crop. To discuss the water needs of the cover crop for its emergence, figure 3(c) shows the mean cumulative precipitation during the months following the starting date of the fallow period.

3.3. Albedo increase and radiative forcing

The strategy to introduce a cover crop in the crop rotation is described in section 2.1.3. In the first scenario that we tested, the cover crop was added, when possible, during the first three months after the harvest of the winter crop (see figure 2(a)). The impact is expressed after averaging the yearly results over three years. The increase in the surface albedo resulting from the introduction of a cover crop (α_CI−α, equation 2) is shown in figure 3(b). The increase in the surface albedo varies geographically and is, on average, equal to 0.0023 over the entire domain during the introduction period. In some places (e.g. France and Romania), the surface albedo of croplands is increased up to 0.15. The magnitude of this increase depends on how dark the soil is (low value of the bare soil albedo, α_bs) and how developed the cover crop is (high value of the vegetation fraction of the cover crop, veg_CI). In a few places, the bare soil is brighter than the cover crop, thus making the introduction of a cover crop unprofitable (see negative values in figure 3(b)). These areas are not discarded from the calculation of the climate change mitigation power. The bare soil albedo has a seasonal cycle, and it usually decreases rapidly after the summer (beginning of the rainy season). During the autumn-winter period, the values of the bare soil albedo across Europe become low (see figures 1(a) and (b)). The introduction of the cover crop becomes potentially profitable, as can be observed in figure 3(b).

Table 2. Mean average increase in the surface albedo and the associated radiative forcing (in W.m⁻² and in MtCO₂-eq.year⁻¹, respectively) per country for 3 month cover crop introduction period. The percentage of the surface used is indicated relative to the total surface of the country (in %). GHG agricultural emissions are listed in the last column on the right (in MtCO₂-eq).

	Surface albedo increase	Albedo-induced radiative forcing	Country area	% use of the country's area for cover crop introduction	Radiative forcing per country (method 1–constant AF)	Radiative forcing per country (method 2–GWP)	Agricultural GHG emissions (2011)
		W.m⁻²	km²	%	MtCO₂-eq.year⁻¹	MtCO₂-eq.year⁻¹	MtCO₂-eq
Portugal	2.70 E–03	−5.88 E–01	90 608	2.70	7.83 E–02	7.24 E–02	30.09
Italy	1.80 E–03	−2.00 E–01	306 697	3.51	21.62 E–02	20 E–02	30.86
Malta	1.64 E–03	−3.46 E–01	350	3.31	0.02 E–02	0.02 E–02	0.06
Sweden	6.85 E–03	−1.06 E–02	449 857	0.12	0.23 E–02	0.21 E–02	7.17
Finland	1.33 E–03	−1.45 E–02	337 050	0.20	0.29 E–02	0.27 E–02	6.41
Estonia	1.23 E–03	−1.13 E–02	45 285	0.48	0.08 E–02	0.07 E–02	1.22
Latvia	2.27 E–03	−2.33 E–02	65 115	0.98	0.25 E–02	0.23 E–02	2.40
Lithuania	2.14 E–03	−2.34 E–02	65 234	1.96	0.56 E–02	0.52 E–02	4.35
Poland	2.84 E–03	−8.74 E–02	315 623	5.40	15.19 E–02	14.05 E–02	30.09
Czech Republic	2.81 E–03	−1.04 E–01	79 686	5.39	5.21 E–02	4.81 E–02	7.90
Slovakia	5.06 E–03	−1.88 E–01	49 267	7.10	4.83 E–02	4.47 E–02	2.81
Spain	5.42 E–04	−1.04 E–01	507 174	2.99	17.31 E–02	16.01 E–02	34.24
Austria	4.79 E–03	−2.18 E–01	85 066	3.84	4.86 E–02	4.50 E–02	7.15
Hungary	5.36 E–03	−1.71 E–01	94 490	13.09	13.56 E–02	12.54 E–02	5.88
Romania	6.10 E–03	−2.27 E–01	243 099	10.62	36.05 E–02	33.34 E–02	17.77
Slovenia	6.14 E–03	−2.35 E–01	20 777	2.72	0.74 E–02	0.68 E–02	1.70
Croatia	4.74 E–03	−1.64 E–01	56 471	4.89	1.85 E–02	1.71 E–02	2.80
Bulgaria	6.38 E–03	−3.73 E–01	112 391	9.47	38.87 E–02	35.95 E–02	4.90
Greece	2.80 E–03	−3.09 E–01	136 016	5.30	12.80 E–02	11.84 E–02	8.57
Cyprus	3.41 E–04	−8.05 E–02	9 690	0.52	0.11 E–02	0.10 E-02	0.62
France	3.30 E–03	−2.96 E–01	557 639	7.87	87.38 E–02	80.82 E–02	77.36
Ireland	1.31 E–03	−7.03 E–02	70 604	0.81	0.52 E–02	0.48 E–02	17.75
United Kingdom	1.04 E–03	−5.42 E–02	246 934	1.66	5.05 E–02	4.67 E–02	44.01
Belgium	4.38 E–03	−1.58 E–01	30 714	8.90	2.38 E–02	2.20 E–02	10.14
Netherlands	3.04 E–03	−7.03 E–02	37 571	5.08	1.97 E–02	1.82 E–02	18.17
Luxembourg	2.81 E–03	−1.08 E–01	2 721	3.77	0.04 E–02	0.04 E–02	0.66
Germany	2.81 E–03	−1.47 E–01	362 390	6.20	35.14 E–02	32.5 E–02	64.54
Denmark	1.22 E–03	−4.49 E–02	43 696	3.97	1.33 E–02	1.23 E–02	10.33
Europe (EU-28)	2.48 E–03	−1.49 E–01	4 422 217	4.49	3.16	2.92	426.28

Data from the ECMWF reanalysis are used. Figure 3(a) shows the mean incoming solar radiation multiplied by the atmospheric transmittance (T_a⁎ SW_in, equation 2) during the first three months of the 2008–2009 fallow period (the same period during which the difference in the albedo was calculated in figure 3(b)). This number represents the outgoing solar radiation from the Earth in the case of a surface albedo equal to 1. The mean value is 39.26 W.m⁻², and geographically, it primarily depends on the mean solar zenithal angle, which explains the south–north gradient. Figure 3(c) shows the change in RF_Δα in W.m⁻² caused by the introduction of a cover crop after the winter crop harvest in 2008 (the result of a combination of figures 3(a) and (b), see equation (2). The areas with a strong potential are not always the areas that exhibit the strongest albedo increase (see, for example, southern Portugal in figure 2(b)), as high values of atmospheric transmittance and incoming solar radiation are also important. However, the strongest changes in RF_Δα occur in Romania, Bulgaria, Hungary, Slovenia, and northern France, and these variations are clearly caused by strong albedo increases.

We estimate that the areas where cover crops could be introduced cover 4.17% of the 28 member states of Europe, including 7.87% of France, 9.47% of Bulgaria, 10.62% of Romania, and 6.20% of Germany (see table 2). This area represents 22% of the cultivated crop areas in the EU-28 (34% in France, 30% in Romania, and 21% in Germany). The difference in radiative forcing is lower for high latitude regions where less incoming solar radiation reaches the surface at the time of the cover crop introduction (0.01 W.m⁻² in Sweden versus 0.59 W.m⁻² in Portugal on average).

Table 2 also lists the cover crop albedo-induced effect on the radiative forcing per country in the EU-28 converted into MtCO₂-eq.year⁻¹ by using method 1 (and method 2, see section 2.2.5). France, Bulgaria, Romania, and Germany are the four countries with the greatest mitigation potentials, with changes in RF_CO₂ of 0.87 (0.81), 0.39 (0.36), 0.36 (0.33) and 0.35 (0.33) MtCO₂-eq.year⁻¹, respectively. These values are consistent with the crop yields at the national scale, as France is by far the most important agricultural producer of cereals in Europe, according to Eurostat. The cumulative RF_CO₂ over EU-28 is 3.16 (2.92) MtCO₂-eq.year⁻¹. This value represents 0.74 (0.68)% of the agricultural greenhouse gas (GHG) emissions in 2011, which were equal to 426.28 MtCO₂-eq (The Eurostat statistics were updated in 2015, http://ec.europa.eu/eurostat/) (see table 2).

To go further in our study, we tested a second scenario, still based upon a 3 month cover crop development, but accounting for limitations due to the water requirements of cover crops (see section 2.2.6 and figure 2(c)). By doing this, we estimated that, for EU-28, the water requirements for the emergence of cover crops could decrease the albedo effect by 28% in the case of no irrigation. The resulting cumulative RF_CO₂ over EU-28 was 2.27 (2.10) MtCO₂-eq.year⁻¹.

Finally, in the third scenario, we calculated the greatest impact of extending the cover crop for a period longer than 3 months, i.e. up to a maximum of 6 months, depending on the duration of the fallow period for each pixel (for that, no water limitation was taken into account). The cumulative RF_CO₂ over EU-28 was 4.31 (3.99) MtCO₂-eq.year⁻¹. With this extension of the cover crop life, we estimated that a compensation level of up to 1.01 (0.93)% of the agricultural GHG emissions could be obtained for the EU-28.

4. Discussion and conclusions

Potential level of climate mitigation. By using MODIS satellite data and the approach developed by Carrer et al (2014), we were able to identify areas where the soil is dark enough (low albedo value) to make the introduction of a cover crop profitable. We examined the most likely scenario by introducing cover crops in crop rotations (between the winter and summer crops). We estimated that cover crops could potentially be introduced for 3 months over 4.17% of the surface area of the EU-28. That would represent 22% of the European cultivated surface area. This value at the European spatial scale is quite consistent with the estimate of Poeplau and Don (2015). They estimated that 25% of the global cropland areas (16 million km²; Siebert et al 2010) could potentially be cover cropped.

In Europe, we show that the mean average increase in the surface albedo over these cultivated areas and the associated RF_Δα of these cover crops are 0.0025 and −0.149 W.m⁻², respectively, for the 3 month scenario. If we convert this RF_Δα into CO₂-eq, it corresponds to a mitigation potential of 3.16 MtCO₂-eq.year⁻¹, assuming that the cover-crop practice is maintained for a period of 100 years (2.92 MtCO₂-eq.year⁻¹, according to the GWP method). Based on our current knowledge, the estimated uncertainty due to the CO₂ equivalent conversion methodology is approximately 8%. However, it should be remembered that long analysis times, such as that of the current study, deemphasize the RF_Δα expressed in CO₂ equivalent (Anderson-Teixeira et al 2012). Nevertheless, this mitigation potential would represent, every year for a 100 year period, 0.74% of the of the EU-28 agricultural GHG emissions in 2011, which is equal to 426.28 MtCO₂-eq. In other words, the introduction of cover crops for 100 years would compensate for 74% of the human-induced GHG agricultural emissions of one year (based on 2011). France, Bulgaria, Romania, and Germany appear to be the four countries with the greatest potential. We show that Bulgaria and Romania have chernozem soil with very low surface albedo in the winter (figure 1). This explains why those two countries are among the countries with the greatest mitigation potentials. The mitigation potential is also probably important for neighbouring countries outside Europe that also have chernozem soil (Moldavia, Ukraine, and Russia, in which the percentage of land used for agriculture is very high).

This mitigation potential of 3.16 MtCO₂-eq. year⁻¹ is equivalent to a mitigation potential per unit (m²) of an introduced cover crop over Europe of 15.91 gCO₂.year⁻¹.m⁻². We show that this mitigation potential could be increased by 27% if the cover crop is extended for periods longer than 3 months. This magnitude order is consistent with the recent estimations of the albedo effect delivered by (Kaye and Quemada 2017). They showed, using case study sites in central Spain and Pennsylvania (USA), that the surface albedo change due to cover cropping may mitigate 12–46 gCO₂.year⁻¹.m⁻² over a 100 year time horizon (impact per m² of cover crop). Additionally, they estimated that the increase in soil carbon sequestration rates and the decrease in fertilizer due to the adoption of the cover crop practice should mitigate greenhouse gas-based climate change by ~116 gCO₂.year⁻¹.m⁻² for non-legumes and by ~135 gCO₂.year⁻¹.m⁻² for legumes. This result is consistent with the value of 110 gCO₂.year⁻¹.m⁻² (per surface unit of cover crop) of the soil C storage effect found by Poeplau and Don (2015). Considering the net biogeochemical effects of cover crops found in Kaye and Quemada (2017) and adding the albedo effect found in this study, the total cover crop mitigation effect would be close to ~150 gCO₂.year⁻¹.m⁻², which would result into a cumulative value over Europe of 29.79 MtCO₂-eq.year⁻¹ (considering that cover crops could be introduced to 4.17% of the EU-28, as explained below). As the GHG agricultural emissions in 2011 were equal to 426.28 MtCO₂-eq., we believe that the introduction of cover crops over Europe may mitigate up to 7% of the human-induced GHG agricultural emissions per year, considering 2011 as the reference year; the impact of the albedo change per year would be between 10% and 13% of this total impact.

Advantages and limitations of cover crop adoption. Little is known about the impact of the introduction of cover crops at a large scale on the climate, even if this practice tends to be imposed on farmers by European legislation for its nitrogen-capture effect. Hence, it is unclear if these changes will generate climate feedback (such as changes in cloudiness resulting from other biogeophysical impacts such as disturbance of roughness or evapotranspiration). According to Ceschia et al (2017), the albedo cooling effect of cover crops could be doubled when considering their effect on long-wave radiation. Indeed, in a field scale comparative experiment (cover crop vs bare soil), they showed that the dynamic and intensity of the longwave effect was very similar to the albedo-induced cooling effect. Additionally, they found that the radiative cooling effects would be reinforced by a decrease/increase in the sensible/latent heat fluxes at the surface. In another study, Tribouillois et al (2018), showed that, compared to bare soil, cover crops increased evapotranspiration (i.e. latent heat fluxes) without limiting the water resources for the next crop, if the cover crop were buried one month before seeding.

Additionally, reducing the area of bare soils by sowing cover crops provides a number of additional ecosystem services (e.g. reductions in soil erosion and nitrogen leaching and increases in biodiversity and soil fertility (Justes et al 2012) and probably has an impact on the proliferation of weeds, pests or pathogens. Furthermore, benefits in terms of reducing the use of fertilizers (and associated emissions) should be considered when the cover crops are leguminous. Some disadvantages and limitations to such practices also exist. Although the long fallow period may allow the timing of the cover crop introduction to be adjusted, we showed that the water requirements for the emergence of cover crops could decrease the albedo effect by 28% without irrigation. Furthermore, there is a short-term extra financial cost to farmers associated with cover crop cultivation, as well as small additional GHG emissions caused by seeding and destroying the cover crops (however, these emissions are small compared to the C storage benefits) (Ceschia et al 2010). All these feedbacks mechanisms must be carefully evaluated in the upcoming by using climate, economic, and ecological coupled models. From our perspective, a financial compensation for this climate change mitigation service should be encouraged (e.g. via a carbon market). In addition, the cultivation of crops with high water needs should be avoided if it is found that the effect of an additional cover crop would exceed an optimal limit, which still needs to be defined in accordance with equivalent economic criteria. In future work, feedback mechanisms should be analysed to refine the benefits proposed in this study.

Following the COP21. Evidence currently exists that the global warming issue needs to be resolved using attenuation measures and that limitations on GHG emissions are no longer sufficient. Following the COP21 meeting in 2015, the target is to limit global warming below a threshold of 1.5 °C. This new directive, which replaces the previous threshold of 2 °C, has great consequences, as it cannot be achieved even if we were to immediately stop all GHG emissions. Consequently, the resulting outcome of the COP21 is that policymakers will likely need to use geoengineering services for climate mitigation. Using satellite data, this study is in line with other recent studies that estimated the potential mitigation effect of cover cropping. The overall potential of mitigation (resulting from the change in albedo, carbon sequestration, and the change in fertilizer use) could be an appropriate solution (or contribution) that should be encouraged through agricultural policies in the future. As a matter of fact, the introduction of cover crops is in line with the on-going reform of the EU's Common Agricultural Policy, which aims to foster a greening of agricultural surfaces to fight against climate change and for ecological matters. In the upcoming years, the approach presented in this article may, therefore, be weighted against other geoengineering strategies, such as sulfate injections into the atmosphere, according to scientific and ethical criteria.

Acknowledgments

This work was supported by the GHG-Europe (FP7 European grant no. 244122) and Fluxpyr (FEDER Interreg IVa program, ref POCTEFA 08/34) research programs, ADEME (REACCTIF program through the CICC project), Agence de l'Eau Adour-Garonne (BAG'AGES project) CNRS-INSU who is supporting the OSR (Observatoire Spatial Regional), and the Région Midi-Pyrénées Council, and the LSA-SAF project of EUMETSAT. The authors thank the MODIS team who are developing satellite-derived surface albedo products (especially Qingsong Sun and Crystal Schaaf) and the teams that are maintaining the Lamasquère site. The authors also thank J L Roujean, N Boukachaba, E Azzi, C Planque, S Faroux, and R Bright for their assistance.

What is the potential of cropland albedo management in the fight against global warming? A case study based on the use of cover crops

Article metrics

Submit

Author e-mails

Author affiliations

ORCID iDs

Dates

Peer review information

Abstract

1. Introduction