Soil-Surface Straw Influences Micrometeorological Conditions Affecting Canola Mortality During Nighttime Frosts

Our objective was to measure alterations in the micrometeorological conditions surrounding canola seedlings during frost periods, and to quantify seedling mortality as a function of straw distribution on the ground surface. The data was acquired from 15 frosts in 2014. We used four treatments, comprising ground surface without straw (SWS), ground surface entirely straw-covered (SEC), sowing line without straw (SLW), and soil with preexisting surface straw (SES), over three experiments. Net radiation (NR), soil heat flux (G), air (Ta), leaf (Lf), rosette (Tr), and surface temperature (Ts), and plant mortality were evaluated. NR was higher in the SEC treatment and lower in the SLW treatment, whereas G was higher on straw-covered ground; Ts and Ta were lower in the SEC than in the other treatments during the most intense frosts. On 06/19, Tr in the SEC and SLW treatments was -0.66 °C and 0.42 °C, respectively; on 08/14, Lf was -3.62 °C and -2.88 °C in the SEC and SLW treatments, respectively. Plant mortality due to the frost on 06/19 was 30% in the SEC treatment, but 0% in the SLW treatment; the frost of 08/14 caused 33.8% mortality in the SEC treatment and 1.25% in the SLW treatment. This therefore showed that removing straw from the sowing line improved the microclimate around the plants, thus reducing canola mortality at the beginning of the growth cycle, which is when frost events most frequently occur.

other field crops has been obtained using artificial frost, which is limited in its ability to reflect natural frost conditions (Frederiks et al., 2012;Martino & Abbate, 2019).
Given the importance of no-tillage farming systems in Brazil, together with the prevalence of frosts that damage canola seedlings, a better understanding of the influence of the soil surface type on micrometeorological variables and the response of canola plants to cold is required. Quantifying micrometeorological variables in the soil-plant-atmosphere interface during periods of frost can improve the agroclimatic zoning of canola, and aid in the establishment of growth and development models.
Thus, the objective of this study was to identify and quantify alterations in the micrometeorological conditions surrounding canola seedlings during periods of frost, and to quantify seedling mortality as a function of the distribution of straw on the ground surface.

Study Location
In 2014, the effects of frost events (n = 15) on canola were evaluated in three experiments, under field conditions, in the research area of the Embrapa Trigo, in Coxilha, RS,Brazil (28°10′S,52°19′W;altitude,686 m). According to Köppen's classification, the climate of the region is Cfa (subtropical wet with warm summer) (Alvares et al., 2013). The soil type in the research area is Rhodic Hapludox. The average declivity of about of 3-5% is similar to most grain fields in the southern region of Brazil.

Experimental Design
The three experiments were sown on April 23 (experiment 1), May 29 (experiment 2), and July 16 (experiment 3), 2014, these dates being within the period when frosts were most likely to occur in the region (May to August) (Alvares et al., 2013). The experimental design was a complete randomized block with four replicates. Each experimental unit (area, 30 m²; length, 6 m), contained 16 rows of canola plants oriented in an E-W direction along the width of the unit. The treatments consisted of four straw distribution patterns: ground surface without straw (SWS), ground surface entirely covered with straw (SEC), sowing line without straw (SLW), and soil with preexisting surface straw (SES). Fertilization was based on the canola requirements and soil analysis. Weeds and insects were controlled in accordance with standard guidelines for canola (Thomas, 2003).
There were preexisting straw residues of 4300 kg ha -1 at the soil surface; this was therefore the quantity present in the SES treatment, and disruption of the soil covering was generated solely by the passage of the seeder. In the SWS treatment, preexisting residues were completely removed from the soil surface. In the SEC treatment, 3000 kg ha -1 of soybean straw was distributed on top of the preexisting crop residues, therefore totaling 7300 kg ha -1 . Finally, in the SLW treatment, straw was removed to a distance of approximately 5 cm from each side of the sowing lines and repositioned between the lines. In addition, 1500 kg ha -1 of soybean straw was added in order to replicate the quantity of straw and extent of surface coverage in the SEC treatment.

Measurements and Analysis
Evaluations of both the micrometeorological and plant variables were performed up to the D2 stage (bud covered with visible secondary inflorescences) of the canola, based on the phenological scale proposed by the Centre Technique Interprofessionnel des Oléagineux et du Chanvre (CETIOM). Micrometeorological monitoring was only conducted in the first and second experiments (one experimental unit per treatment), owing to the limited availability of equipment and sensors.
an angle of 45° relative to the ground, oriented so that its reading area would cover approximately 0.5 m 2 . All sensors (Copper-Constantan Type T, HC2S3 capacitive sensor, and SI-111 infrared thermometry sensor) were positioned along one of the plant lines.
Leaf (Lf) and rosette (Tr) temperatures were measured using copper-constantan thermocouples (thickness, 0.15 mm; Type T). Sensors thermocouples were installed inside the main veins of developed and fully exposed leaves was measured. Tr was measured in the second experiment, between June 18 and June 25, because plants had one leaf during that frost. For these evaluations, three sensors were installed per treatment in different canola plants, located in the center of the experimental unit, totaling 12 sensors per frost. In the third experiment, Lf measurements were instead made with a portable infrared thermometer (UX-40, Ircon, New Dehli, India) (ε = 0.98). Evaluations were performed on three plants (three replicates) located in the center of the experimental unit per treatment. The first reading was taken at 04:00 and then repeated every half hour until 08:00.
Measurements of NR, G, Ta0.03, Ta0.27, Ts, and Lf were made every 30 s, with averages calculated every 15 min. Measurements of U and UD were taken every 30 s. Data were stored in a datalogger (CR 1000, Campbell Scientific, Loughborough, UK).
On the night of 05/24, fog formed when U was above 1 m s -1 (Figure 1), a level that favors the mixing of stratified cold air layers, promoting condensation of excess vapor (Nemitz, 2009). The U value was above 1 m s -1 from 23:10 to 00:20, resulting in an increase of 2.12 °C for Ta0.27. Thus, both Ta0.27 and Tdp remained the same until 00:20 ( Figure 1); this was a necessary condition for fog formation, which reduced the loss of radiation energy from the ground surface, thus attenuating cooling and maintaining NR at near zero ( Figure 1). When U fell below 1 m s -1 , Tdp and Ta0.27 also reduced. Thus, Ta0.27 was slightly higher than Tdp (00:00 to 02:30), thus reducing fog formation. After 02:30, Ta0.27 returned to the same value as Tdp, but the NR data allowed us to infer that there was no longer any fog.   Vol. 12, No. 11; these frosts, however, the difference between Ta0.27 and Tdp was lower than on 08/14 because the air vapor content was higher in comparison (Figures 2, 3, and 4).

Plant Mortality, and Temperature of Leaves and Rosettes
In the SEC treatment, canola plant mortality was 30% and 33.8% after the frost events of 06/19 and 08/14, respectively ( Table 2). The lower thermal conductivity of soil covered with straw (0.18 W m -1 K) when compared with uncovered clay soil (0.56 W m -1 K) (Ahn et al., 2009) reduced the energy exchange between the soil and atmosphere. This resulted in greater cooling in straw-covered soil, resulting in the higher canola plant mortality observed in the SEC treatment compared with the other treatments (Kaspar & Erbach, 1998;Cierniewski et al., 2015;O'Brien & Daigh, 2019). Note. CV, coefficient of variation; Averages followed by the same lower case letter indicate there are no significant differences (Tukey's test, p > 5%) at each frost date.
from the plants. In the SEC treatment, Tr was lower than in the other treatments (Figure 2), due to the greater distance from the ground surface and direct contact with the straw.
Canola plants in the third experiment had three developed leaves by the frost of 08/14. From 04:00 to 06:30, Lf was lower in SEC than in the other treatments (Table 3), which resulted in a higher plant mortality (Table 2). NR and Ts increased at sunrise (07:02), followed by Ta0.03 and Lf (Figure 4 and Table 3). Thus, at 07:00, there was no difference in Lf between the SEC and SES treatments (Table 3). From 07:00 to 07:30, Lf fell in the SWS treatment, remained stable in the SLW treatment, and increased in both the SEC and SES treatments ( Table 3). The straw on the soil surface warmed more quickly than the bare soil surface because of the low thermal conductivity, low specific heat, and greater degree of aerodynamic roughness (Van Doren & Allmaras, 1978;Horton et al., 1996;Ahn et al., 2009;Lal & Stewart, 2012). This resulted in a larger increase in Ta0.03 and Lf in SEC compared with the other treatments.  (Table 3). According to Pearce (2001), Bredow and Walker (2017), Palta and Weiss (2018), Takahashi et al. (2019), Hincha and Zuther (2020), Ramlov and Friss (2020), the main cause of frost-related damage is the formation of extracellular ice crystals, which have a lower vapor pressure relative to liquid water in the cytoplasm, and promote the intracellular transfer of water to intercellular spaces. The rapid increase in Lf on frosty mornings precipitates thawing and exacerbates the potential for cellular damage. Slow thawing processes favor gradual hydration of cell membranes and, consequently, cell recovery. Straw on the soil surface does not assist in this process, as it accelerates warming early in the morning and exacerbates cooling at night, both of which increase the potential for frost damage in plants.
The leaf area index (LAI) was negligible on 06/19, as the canola plants were at phenological stage B1 (one developed leaf). On 08/14, when canola plants were at phenological stage D2 (bud covered with visible secondary inflorescences), the LAI was 3.25, 2.38, 2.30, and 2.02 in the SWS, SES, SLW, and SEC treatments, respectively. Thus, the LAI may have contributed to reducing the difference in surface radiative energy loss between the frosts of 08/14 and 06/19. However, the amount and distribution of straw on the soil surface influenced the soil-air energy balance during frost events. A straw surface, such as in the SEC treatment, can reduce long-term energy loss by the emission of longwave radiation when compared with conditions such as in the bare soil (SWS), SES, and SLW treatments. Our results corroborated the observations of Cierniewski et al. (2015) and O'Brien and Daigh (2019), who suggested that straw acts as an insulator by restricting radiative exchange between the soil and atmosphere. Thus, surfaces with the largest amounts of straw on the soil surface displayed a lower energy loss from longwave radiation emission because they exhibit lower temperature values, according to the Stefan-Boltzmann law, which states that the energy radiated by a surface is proportional to the fourth power of its temperature (Snyder & Melo Abreu, 2010). Azooz et al. (1997) found that in a soil with a 30 cm residue-free strip over the center of the row had significantly higher heat flow into the soil surface. These result indicated that a residue-free strip over the row center alternated with residue strip in between rows could be important in providing more heat input into the soil surface at the row center. Considering the NR results, the removal of straw from the sowing line (SLW treatment) promoted similar responses to a soil surface without straw (SWS treatment), which states that the energy radiated by a surface is proportional to the fourth power of its temperature.

Surface Temperature
The NR data indicated that ground surfaces in the SEC treatment were colder than in the other treatments ( Figures  1, 2, 3, and 4); this was confirmed by the average Ts data for 06/19 (2.09 °C, 2.89 °C, 2.93 °C, and 3.00 °C in the SEC, SWS, SES, and SLW treatments, respectively) ( Figure 2). However, Ts values did not correspond to NR values on 08/14 because of the lower Ts values observed in the SWS treatment (Figure 4), probably due to LAI interference. In addition, the minimum Ta0.03 (below leaf level) was -2.00 °C, while Ta0.27 in the plant canopy was -3.72 °C. This difference supports the hypothesis that LAI interfered with nighttime NR by acting as a slight barrier that served to reduce radiative loss from the soil surface, thus slowing cooling in the lower third of the canopy.
On 05/24, higher NR and lower Ts values were recorded in the SES treatment (Figure 1), corroborating the Stefan-Boltzmann law (Snyder & Melo Abreu, 2010). The occurrence of lower Ts values in the SES compared with the SEC treatment was probably due to the uneven distribution of straw on the soil surface in the former on 05/24, which reduced energy loss from the ground (Figure 1).
On 07/26, NR values were lower in the SES compared with other treatments (Figure 3). The average NR values from 18:00 to 08:00 were -85.2 W m -2 , -103.7 W m -2 , -96.3 W m -2 , and -96.7 W m -2 in the SEC, SLW, SES, and SWS treatments, respectively ( Figure 3). An exception occurred between 18:45 and 21:30, when the SEC treatment exhibited lower NR values than in the other treatments ( Figure 3). Lower Ts values also occurred in the SEC treatment from 18:45 to 21:30, therefore showing a trend consistent with NR ( Figure 3). The average Ts values were 3.73 °C, 5.02 °C, 5.03 °C, and 5.55 °C in the SEC, SLW, SES, and SWS treatments, respectively, and higher NR values occurred in the SEC treatment. From 21:30 to 08:00, the SWS treatment showed lower NR values, while lower Ts values occurred in the SEC treatment. This contradictory response of NR to Ts is probably related to interference by energy advection.
On 07/26, from 20:10 to 20:30, Ts values increased by 1.10 °C, 0.51 °C, 0.50 °C, and 0.28 °C in the SEC, SLW, SWS, and SES treatments, respectively. This increase coincided with the increase in wind speed from 0.2 m s -1 to more than 1.0 m s -1 (Figure 3). This agreed with the hypothesis that energy advection occurred in the experimental area and/or forced air mixing occurred, changing the air temperature profile.
From 21:00 to 21:10, the wind speed was below 0.5 m s -1 , and Ts decreased by 0.99 °C, 0.56 °C, 0.49 °C, and 0.38 °C in the SEC, SLW, SES, and SWS treatments, respectively (Figure 3). The reduction in Ts in the SEC treatment during this time period highlights the effects of soil surface mulching on reducing Ts on frost nights. In addition, the response of Ts values from 20:10 to 20:30 demonstrates how the energy advection effect (and/or forced air mixing changing the air temperature profile) can minimize Ts differences between straw-covered and bare soil surfaces, because of attenuation of cooling of the straw-covered soil surface. Energy advection in the surrounding areas (and/or homogenization of the air temperature profile) was the cause of the inconsistencies between NR and Ts values, the reasons for which could not be determined conclusively. However, oscillations in NR followed variations in wind speed, which is consistent with evidence that changes in wind speed directly affect NR. Fluctuations in NR values in response to wind speed changes were observed on all frost dates; on 05/24 and 07/26 (low-intensity frost events), oscillations in NR values occurred in all the treatments (Figures 1 and 3), but occurred only in the SWS, SES, and SLW treatments on 06/19 and 08/14 (the two most intense frost events) (Figures 2 and 4).
On 06/19 and 08/14, NR values in the SEC treatment remained more or less constant over most of the monitored period and were lower than in the other treatments (Figures 2 and 4). The exceptions were between 20:30 and 23:30 on 06/19 and 18:00 and 03:00 on 08/14, when the SWS treatment showed similar NR values to the SEC treatment (Figures 2 and 4).This response shown in the SEC treatment was probably due to the greater cold intensity on 06/19 and 08/14.
Parity of measured values between the SEC and SWS treatments prevailed when the wind speed exceeded 1 ms -1 , a speed considered by Nemitz (2009) to be sufficient to promote turbulence in the plant canopy. This may account for the similarity in NR values between these two treatments, given that it cannot be explained by their Ts values.
Thus, the variables measured on the frost dates indicated that the form of soil-surface cover influenced NR (and consequently the energy balance), and also therefore the leaf, surface, and air temperatures. In addition, local topography favored energy advective transport from surrounding areas, which prevented a more pronounced freezing. Interference by energy advection may have lessened the effect of surface straw quantity and distribution on energy balance, and thus on the leaves, surfaces, and air, since lower temperatures occur when there is more straw. However, even taking energy advection into account, significant plant mortality occurred during frost events in treatments with more soil surface straw cover around cold-susceptible canola seedlings with no previous period of acclimation. Thus, due to the energy advection observed in this experiment, it can be inferred that soil surface straw should cause far more damage to young canola plants than in this experiment. Therefore, the removal of straw cover from the sowing line is an effective alternative to reducing frost damage on canola plants.

Conclusions
The results of the experiment showed that removing straw cover from the seeding line can reduce frost-related mortality in canola plants prior to developing three leaves (B3 stage). The air temperature near the surface of leaves and rosettes is lower on completely straw-covered soil than on soil without surface straw. Removing straw cover from 0.05 m on each side of the canola seeding line (i.e., a total of 0.10 m) has a similar effect to the total removal of straw from the soil surface. An increase in wind speed at night associated with the occurrence of frost favors energy advection, and consequently attenuation of cooling.