Introduction

Soil temperature is critical for crop production because it affects seed germination, seedling emergence, root growth, nutrient uptake, and plant development. Soil temperature is also environmentally important because it controls soil microbial communities that are closely linked to ecosystem functioning, such as greenhouse gas generation and carbon and nitrogen dynamics. A positive correlation between soil respiration and soil temperature is well established (Singh and Gupta 1977; Reich and Schlesinger 1992), and studies overwhelmingly show that warmer soils increase the decomposition of soil organic matter (Jenkinson and Ayanaba 1977; Lloyd and Taylor 1994; Trumbore et al. 1996; Dalias et al. 2001; Davidson and Janssens 2006; Sierra et al. 2015; Webb et al. 2016). Hence, elevated soil temperatures can exacerbate soil carbon dioxide efflux and global warming.

Soil temperature is affected by many factors including meteorological conditions, soil topography, soil water content, and plant canopy cover (Paul et al. 2004). No-tillage and cover cropping may change soil temperatures through the accumulation of crop residues and increased shading. Based on 11 cover crop versus no cover crop comparisons, Blanco-Canqui and Ruis (2020) concluded that soils under cover crops are cooler during the day and warmer at night and also warmer during winter and cooler during the rest of the year. Blanco-Canqui and Ruis (2020) further found that the difference in soil temperature between cover crops and no cover crops was larger during the spring than during the rest of the year, and that average springtime soil temperature was about 1 °C lower under cover crops. Lombardozzi et al. (2018) found that cover crops in central North America can increase wintertime temperature under snow by as much as 3 °C. Although a temperature increase from 25 to 26 °C increases soil respiration rate by only about 5%, an increase from 0 to 1 °C increases soil respiration by 22% (Lloyd and Taylor 1994). Hence, cover crops can potentially contribute appreciably to greenhouse gas emissions from soil, especially during the overwinter and early spring periods.

The impacts of land management on soil temperature have long been studied because a 1 °C difference in soil temperature can significantly affect crop growth, especially corn (Walker 1969; Barlow et al. 1977). An emerging issue with cover crops is that they can change soil temperatures relative to bare ground, which may in turn impact both crop productivity and global warming (Lombardozzi et al. 2018). Agricultural practices can affect the biogeophysical properties of the land surface by changing latent heat flux and albedo, which may in turn impact climate. For example, cover crops can increase latent heat flux, transpiration, and albedo compared to bare, snow-free land (Kaye and Quemada 2017). In addition, soil temperature is considered a key driver in soil organic carbon (SOC) dynamics in almost all SOC models (e.g., Campbell and Paustian 2015). Gregorich et al. (2017) demonstrated for temperate climates across southern Canada that soil temperature was the single most important factor controlling residue decay rates, easily superseding other soil properties, such as texture, cation exchange capacity, pH, and moisture. Their study further found that use of thermal time may advance the development of simplified soil carbon models by amalgamating the influences of many soil and climatic variables, some of which are not easily measured, into a single cumulative parameter based on readily available soil temperature data. For many cases in which soil temperature data are not accessible, an empirical equation for soil temperature based on air temperature could be helpful for estimating temperature effects in SOC models (Zheng et al. 1993).

Cooler soil under crop residue (e.g., no-till) relative to no residue (e.g., moldboard plow) is commonly observed, with the amount of temperature difference depending on the time of year and cropping history (Drury et al. 1999; Larney et al. 2003; Munoz-Romero et al. 2015). Burrows and Larson (1962) found at Ames, Iowa, that each 1.125 Mg·ha⁻¹ of applied crop residue reduced the average soil temperature by 0.2 °C at the 10 cm depth during May and June. In recent years, more growers in Ontario are using cover crops in rotations for multiple benefits including improved soil stability (Dapaah and Vyn 1998), reduced weeds (O’Reilly et al. 2011), and increased overall soil nitrogen content in the following growing season (Vyn et al. 2000; O’Reilly et al. 2012; Coombs et al. 2017; Yang et al. 2019). A recent study in southwest Michigan showed that soil temperature was strongly associated with cover crop species, and that this in turn affected the efflux of carbon dioxide and nitrous oxide from the soil (Nguyen and Kravchenko 2021).

There is a general lack of information, however, on soil temperatures under living cover crops compared to no cover crops (bare ground) in different seasons. Accordingly, we measured soil temperature under three cover crops (crimson clover and white clover mix, CW; hairy vetch, HV; red clover, RC) and a no-cover crop control (CK) from August (after cover crop planting) until the following May (before cover crop termination for corn planting) for two cropping years in a corn–soybean–winter wheat/cover crop rotation in southern Ontario. The objective was to test two hypotheses regarding soil temperatures under live cover crops during the over-winter and spring periods: (i) soil temperatures under live cover crops differ from soil temperatures under no cover crop, and (ii) soil temperature is affected by the amount of live cover crop biomass.

Materials and Methods

This field study was established in the summer of 2017 on a Brookston clay loam, which is a cool-humid poorly drained lacustrine soil (Orthic Humic Gleysol, Canadian Soil Classification; mixed, mesic Typic Argiaquoll, USDA Soil Taxonomy), at the Honourable Eugene F. Whelan Experimental Farm, Agriculture and Agri-Food Canada, Woodslee, Ontario, Canada (42°13′ N; 82°45′ W). The average soil texture in the top 20 cm was 33% sand, 37% silt and 30% clay by weight, and average soil pH was 6.3. The 10-yr mean annual air temperature and mean annual precipitation were 9.4 °C and 877 mm, respectively. As surface slopes were <0.5%, soil erosion was negligible.

The experiment included three adjacent fields under a corn–soybean–winter wheat/cover crop rotation with every crop present each year. Shortly after winter wheat harvest, wheat stubble was disked, packed and legume cover crops were planted using an International grain drill (16 Jul. 2018; 30 Jul. 2019). The treatments included a CW treatment (Trifolium incarnatum L. mixed with Trifolium repens L., seeding rate = 12.5 +12.5 kg·ha⁻¹), an HV treatment (Vicia villosa L. Roth, seeding rate =25 kg ha⁻¹), an RC treatment (Trifolium pratense L., seeding rate = 12 kg·ha⁻¹), and a CK treatment. Each field plot was 9.15 m wide (to allow 12 corn rows) by 30 m long, and the statistical design was a randomized complete block with four replicates. To estimate cover crop biomass accumulation, above-ground plant samples were collected (20 cm × 75 cm area, n = 2 per plot) after the first hard frost (late November), and again in the following spring (early May) just before cover crop termination for corn planting. The biomass was dried (60 °C) for two weeks, and then dry weight was measured. Fall and spring samples were collected because these cover crops are reasonably winter-hardy in southern Ontario (i.e., not winter-killed), and therefore resume growth after the spring thaw.

Soil temperature was measured hourly from cover crop planting (31 July 2018; 15 August 2019) to cover crop termination (5 May 2019; 15 May 2020) using Decagon 5TM/5TE sensors and EM-50 automated data loggers (Decagon Devices Inc., Pullman, WA, USA). The sensors were installed at four depths (15, 30, 45, 60 cm, n = 4 per depth per treatment) by hand-augering access holes, then backfilling. The data loggers were downloaded biweekly. The sensors were calibrated for the soil type in the lab using intact cores, and reported accuracy and resolution are ± 1 and 0.1 °C, respectively (Decagon Devices, Inc., 5TE User’s Manual, 2016). Each reported sensor reading was the average four different sensors, which we assume controls among-sensor errors and variability.

Soil temperature values were averaged on a field, season and monthly basis for each depth by cover crop combination, and also on a daily basis for both the coldest winter month and the month prior to cover crop termination. Regression was used to determine depth-wise relationships between soil temperature and accumulated cover crop biomass in each plot. Analysis of variance was used to determine the significance of treatment main effects via the LSMeans procedure (SAS Institute, Cary, NC).

Results

Average monthly air and soil temperatures

The average monthly air temperature varied substantially during the cover crop growing periods (Figs. 1a, 1b). The lowest average monthly air temperatures were observed in January 2019 (−4.7 °C) and February 2020 (−1.5 °C), and the highest in August 2019 (22.9 °C) and August 2020 (21.6 °C). The average monthly air and soil temperatures correlated well at all depths, with change in soil temperature lagging behind the change in air temperature. In 2018–2019, there were no appreciable differences in average monthly soil temperature among cover crop species and the CK (Fig. 1a). In 2019–2020, however, higher soil temperatures occurred at 15 cm depth under cover crops relative to the CK during the winter months, and lower soil temperatures occurred at all depths under cover crops relative to the CK during May (p < 0.05) (Fig. 1b). Average soil temperature for the entire cover cropping season was similar in both cropping years, and the winter increase in temperature was offset by the decrease during the following spring. There were, however, higher soil temperatures during 2019–2020 than during 2018–2019, which reflected differences in air temperature between the two years (Table 1, Figs. 1a, 1b).

Fig. 1a.

Monthly average soil temperature at different depths and air temperature (1.2 m) from July 2018 to May 2019.

Fig. 1b.

Monthly average soil temperature at different depths and air temperature (1.2 m) from August 2019 to May 2020.

Table 1.

Mean soil temperature (°C) during cover cropping period (August to May).^a

Average daily air and soil temperatures

As the cover crops were seeded in July for both cropping seasons (after wheat harvest), their full canopies were not well established until mid to late September. Hence, cover crop effects on soil temperatures were generally not appreciable until the late fall–winter–early spring periods.

For the coldest month in 2019 (January), air temperature dropped from −7.5 °C on 28 January to −21 °C on 30 January, which caused daily soil temperature at 15 cm depth to decrease from −1.0 °C on 29 January to −7.4 °C on 31 January for CK, and from 0.2 °C at 28 January to −2.8 °C at 31 January for HV (Fig. 2a). The decrease in air temperature did not affect soil temperatures under CW and RC, however, and the 15 cm temperatures under CK and HV returned to previous levels after 3–4 days when air temperature increased. A similar but more muted pattern was evident at 30 cm depth, and no temperature responses occurred at the 45 and 60 cm depths (Fig. 2a). However, CK had a lower temperature from the end of January 2019 through to mid-February relative to CW (Fig. 2a).

Fig. 2a.

Daily soil temperature at different depths and air temperature (1.2 m) from 15 January to 15 February 2019 (the period with the lowest air temperature).

For the coldest month in 2020, air temperature ranged from 4.3 to −11 °C during 15 January to 15 February (Fig. 2b). These higher air temperatures (relative to 2019) resulted in relatively higher soil temperatures (relative to 2019) at all soil depths, and below-zero soil temperature occurred only in CK and only for a few days (21–23 January) (Fig. 2b). Unlike 2019, a sharp decrease in soil temperature occurred in the CK treatments following an air temperature drop; there was no clear drop in soil temperature during the coldest period of 2020, and an increase in soil temperature occurred for all treatments on 3 February 2020. The daily average soil temperature at 15 cm depth remained lower in the CK (0.4 °C) relative to other treatments (1.4–2.0 °C) for about 10 days, starting 21 January. A similar decrease in CK relative to the cover crops occurred at the 30 and 45 cm depths but to a reduced extent. At the 60 cm depth, soil temperature showed no cover crop effects or temporal fluctuations during the coldest month of 2020.

Fig. 2b.

Daily soil temperature at different depths and air temperature (1.2 m) from 15 January to 15 February 2020 (the period with the lowest air temperature).

Variability in daily average soil temperature increased with increasing air temperature at all soil depths in spring, and treatment differences were evident in late April until mid-May for both 2019 and 2020 (Figs. 3a, 3b). On 5 May 2019, the daily average soil temperature at 15 cm depth was lower for CW (11.1 °C) and RC (10.8 °C) than for CK (12.2 °C) and HV (12.1 °C) (Table 2). The same pattern occurred at the other depths but to a lesser degree (Fig. 3a). In 2020, separation in daily soil temperature among treatments started around 25 April, reached a maximum by 3 May for all soil depths, and persisted until the last measuring day, 15 May 2020 (Fig. 3b). On 15 May 2020, the daily average soil temperature at 15 cm was lower in all three cover crop treatments (10.5–11.2 °C) than in the CK (13.5 °C) (Table 2). Lower daily soil temperature under the cover crops relative to CK also occurred at the deeper soil depths, being as much as 1.8–2.4 °C lower than CK on 15 May 2020 (Table 2). It is also worth noting that springtime daily average soil temperatures were periodically lower under HV than under CW and RC, despite all three crops having similar amounts of accumulated biomass. This may be due to greater sunlight reflectance and lower plant-air heat exchange in HV because of its numerous, closely spaced leaves and vine-like growth habit compared to greater-spaced broad leaves and erect to decumbent growth habit of CW and RC.

Fig. 3a.

Daily soil temperature at different depths and air temperature (1.2 m) from 5 April to 5 May 2019, before cover crop termination and planting corn.

Fig. 3b.

Daily soil temperature at different depths and air temperature (1.2 m) from 15 April to 15 May 2020, before cover crop termination and planting corn.

Table 2.

Lowest and highest daily average soil temperatures (°C).

Relationships between spring temperatures and cover crop biomass

Spring soil temperature shortly before cover crop termination decreased linearly with increasing accumulation of cover crop biomass (Mg·ha⁻¹) (Fig. 4). The coefficients of determination (R²) ranged from 0.38 to 0.71; the y-axis intercepts (a, mean soil temperature when cover crop biomass is set zero) were 12.38 °C at 15 cm, 11.32 °C at 30 cm, 10.62 °C at 45 cm, and 9.73 °C at 60 cm; and the slope constants (b, model predict of soil temperature change for 1 unit (Mg·ha⁻¹) cover crop biomass change) were −0.24 to −0.19 for the top three depths, and −0.09 at 60 cm. It indicates that a 1 Mg·ha⁻¹ increase in cover crop biomass corresponded to −0.24 °C change in soil temperature at 15 cm depth, but only −0.09 °C change in soil temperature at 60 cm depth.

Fig. 4.

The relationships between cover crop biomass and soil temperature in spring by terminating cover crop for corn planting.

Discussion

This study suggests that live cover crops and dead crop residues behave similarly with respect to their impacts on soil temperatures. For example, summer soil temperatures under both cover crops and residues are lower and less variable than those of bare soil (Bristow 1988; Blanco-Canqui and Ruis 2020), but higher than bare soil during the winter (Yang et al. 2018). In addition, the amount of surface residue retained influences the soil temperature regime (Larney et al. 2003), and surface temperature variations among residue and tillage treatments are due mainly to change in the amount of residue cover (Gupta et al. 1983; Larney et al. 2003; Blanco-Canqui and Lal 2009). In the study at Ames, Iowa, Burrows and Larson (1962) found that for each 1.125 Mg·ha⁻¹ of crop residue retained, average soil temperature was reduced by 0.2 °C at the 10 cm depth during May and June. Blanco-Canqui and Lal (2009) similarly observed in Ohio that each Mg·ha⁻¹ of retained corn stover reduced the soil temperature by 0.5 °C at the 10 cm depth during June–October. In this study, a 1 Mg·ha⁻¹ increase in live cover crop biomass resulted in a 0.24 °C drop in soil temperature at the 15 cm depth, a 0.19−0.20 °C drop at the 30 and 45 cm depths, and a 0.09 °C drop at the 60 cm depth. Similar patterns likely occur because both cover crops and surface residues increase soil heat capacity by conserving soil moisture (Joyce et al. 2002), shade the soil surface from incident solar radiation, alter canopy heat balance, and lower thermal conductivity compared to bare soil (van Wijk et al. 1959; Horton et al. 1996; Johnson and Lowery 1985).

The results of this study are also consistent with other cover crop studies; notwithstanding the fact that such studies are very few in number. For example, Evans et al. (2016) and Teasdale and Daughtry (1993) found that cover crops can reduce diurnal changes in soil temperature; and Kahimba et al. (2008) reported that soils in the Canadian prairies under a cover crop were warmer in winter and thawed earlier in spring than soils without cover crop. It has also been speculated that cover crops may reduce extreme fluctuations in soil temperature, thereby potentially reducing abrupt fluctuations in wet-dry and freeze–thaw cycles (Blanco-Canqui and Ruis 2020).

Soil temperature varied with year and cover crop species. A substantial drop in soil temperature following a sharp decline in air temperature in late January 2019 occurred for HV, but not for CW and RC. We attribute this to HV winter-kill, given there was substantially less springtime biomass for HV (2.7 Mg·ha⁻¹) than for CW (6.2 Mg·ha⁻¹) and RC (6.5 Mg·ha⁻¹) (Fig. 5). The apparent biomass effect on soil temperature also occurred in spring 2019, when soil temperatures under HV were similar to CK but significantly higher than in CW and RC. The biomass effect on temperature appeared differently in spring 2020, however, with a greater temperature decrease in HV relative to the previous year, which was likely due to a dramatic increase in HV biomass from fall 2019 to spring 2020 (Fig. 5). These results support the Blanco-Canqui and Ruis (2020) postulation that cover crops reduce soil temperature during warm periods by intercepting and reflecting solar radiation, shading the surface, and reducing soil water evaporation; and that they also likely increase soil temperature during cold periods through the mulching–geothermal gradient effect.

Fig. 5.

Aboveground biomass in fall by freeze-up and in spring by spraying termination before corn planting.

Conclusions

Both hypotheses were effectively supported: live cover crops caused soil temperature to be warmer in winter and colder in spring relative to no cover crop; springtime soil temperature changed linearly with amount of accumulated cover crop biomass.

Near-surface soil temperatures differed among cover crop species for both the coldest and warmest conditions during the cover crop growth cycle (i.e., from August of one year to May of the following year). This was likely due to differences in the amounts of cover crop biomass and cover crop species which intercepting and reflecting solar radiation, insulating the soil surface, and reducing soil water evaporation.

Near-surface soil temperature was consistently increased in winter and decreased in spring by red clover and the crimson clover and white clover mix. Hairy vetch, on the other hand, had inconsistent temperature effects, which we attribute to variable biomass accumulation stemming from its susceptibility to winter-kill in southern Ontario.

Although lower spring and summer soil temperatures could assist accumulation (or slow loss) of soil carbon, they could also exacerbate crop diseases and (or) reduce seed germination, seedling emergence, early crop growth, and crop yield. Similarly, warmer soil temperatures in winter could assist overwinter survival of crop pests and diseases as well as survival of cover crops. In addition, the above effects could be enhanced substantially when cover cropping is combined with no-tillage on cool, humid soils. Hence, the temperature effects of cover cropping may be both positive and negative, and should therefore be considered carefully when interpreting soil biochemical processes, or developing agricultural best management practices.