Effects of depth and cover crop treatment on the functioning and diversity of soil microbial communities

**Included results and discussion are from a draft thesis to be defended in October 2022

3. Results

3.1 Fall 2021

              In the fall, all experimental plots contained soybeans at full maturity (stage R8) at the time of sampling.  The average soil moisture across all experimental plots was 21% (Figure 2a).  Across sampling depths soil moisture steeply declined from 0-13 cm (F_1,78 = 7.82, p = 0.0065, slope = -0.0006±0.0002).  Soil moisture did not differ among cover crop treatments (F_3,9 = 0.69, p = 0.5820; Figure 2b).

              The average well color development (AWCD) increased over the 5-day incubation period.  Preliminary analysis of AWCD indicated the largest difference occurred between days 3 and 4 and so day 4 readings were used for all subsequent analyses.

              The number of carbon sources metabolized by the microbial community declined with depth (F_1,79 = 54.35, p < 0.0001, slope = -0.2646±0.03589).  On average, 80% of the 31 carbon sources were metabolized at the soil surface and this average declined to half of the sources used at 43 cm (Figure 3a ).  While richness changed with depth it  did not differ among cover crop treatments (F_3,9 = 0.62, p = 0.6169; Table 1).  Evenness (the similarity in color development among wells) also declined with depth (F_1,79 = 39.24, p < 0.0001, slope = -0.00257±0.000410; Figure 3c) but did not differ among cover crop treatments (F_3,9 = 0.69, p = 0.5792; Table 1).  Because these two components of diversity exhibited trends for depth but not cover crop, a single measure of diversity, Shannon’s H, also declined with depth (F_1,79 = 44.57, p < 0.0001; Figure 3e).   Thus, as depth increased, there were fewer carbon sources being metabolized, more variation in the degree of metabolism, and diminished functional diversity. 

              The first three principal components (PC) from analysis of the EcoPlate™ AWCD explained 74.8% of the  variation in the microbial metabolism of the 31 carbon sources.  Inspection of factor loadings shows high correlation among carbon sources (Supplementary Table 1).  Most of the carbon sources loaded on PC 1 and PC 2,  with 11 loading heavily on both factors (Table 2).  PC 2 was strongly and positively correlated with carbohydrates and polymers, and Factor 3 was positively correlated with carboxylic acids (Table 2).  Analysis of the factors revealed that the cover crop planted the previous season significantly affected both PC 2 (F_3,9 = 4.00, p = 0.0460) and PC 3 (F_3,9 = 8.28, p = 0.0059).  Control plots scored high for PC 2 and low for PC 3, whereas pennycress plots scored low for PC 2 and high for PC 3 (Figure 4a).  This indicates that soil in the control plots included a greater density or diversity of microbes capable of metabolizing a range of carbohydrates and polymers compared to soil from pennycress plots (t_3,9 = 3.41, p = 0.0325; Figure 4a).  Whereas pennycress plots (t_3,9 = -4.63, p = 0.0056) and mixed cover crop plots (t_3,9 = -3.72, p = 0.0203) metabolized a greater amount of carboxylic acids than did control plots.  While the control and pennycress tended to separate out, the cereal rye and the mixed species cover crop clustered together (Figure 4a), indicating more balanced metabolism of carbon sources.  Sampling depth had a significant impact on PC 1 (F_1,79 = 25.46, p < 0.0001, slope = -0.026+0.005), PC 2 (F_1,79 = 24.78, p < 0.0001, slope = -0.028+0.006), and PC 3 (F_1,79 = 4.17, p = 0.0445, slope = -0.012+0.006   As depth increased, values for all PCs declined.

              A closer look at how the metabolism of each functional group of carbon source changed across depths and among cover crop treatments  showed that usage of all functional groups declined significantly with increasing depth (Figure 5, Table 4).  Across all depths, the trend was that phenolic compounds were least utilized, and at shallower depths that polymers were the most utilized.  Beyond depths of 13 cm there were less differences in the utilization of each functional group.  The most differences in functional group usage were observed from the surface to 4 cm.  The cover crop planted the previous season did not have a significant effect on the pattern of functional group usage (Table 3).   

3.2 Spring 2022

              Cover varied among cover crops in the spring and was more heterogeneous than observed for soy in the fall. Because the goal of this research was to identify effects of cover crops on soil microbes, I deliberately chose areas of the experimental plots in the spring where coverage was highest and haphazardly sampled within those areas. Cereal rye provided the greatest coverage (83.5%), followed by pennycress (54.8%) and mixed (48.4%).  Control plots had the least amount of plant coverage (4.04%).  There was overall poor establishment of mixed cover crop experimental plots.  At the time of sampling these plots contained mostly radishes, with few peas, clover, or oats.  Control plots contained a mix of weeds, with high overall plot coverage and I sampled locations in control plots that had little plant material.  The average soil moisture across all plots in the spring was 15% (Figure 2a).  Similar to the fall, soil moisture steeply declined from 0-13 cm (F_1,78 = 10.82, p = 0.0015 Control plots had greater soil moisture than cover crop plots (F_3,9 = 8.43, p = 0.0056) (Figure 2b).

              As with the fall, AWCD increased over the 5-day incubation period.  Preliminary analysis of AWCD indicated the largest difference occurred between days 2 and 3, and no difference was found between days 3 and 4.  Day 3 readings were used for subsequent analyses in the spring. 

              The number of carbon sources metabolized by the microbial community  (richness) declined with depth (F_1,79 = 116.15, p < 0.0001, slope = -0.2279±0.02115; Figure 3b).  On average, 74% of the 31 carbon sources were metabolized at the soil surface and this declined to 30% of the sources used at 43 cm.  Richness did not differ among cover crop treatments (F_3,9 = 0.43, p = 0.7349; Table 1).  Evenness declined with depth, indicating that the community is becoming more variable as depth increases and some microbes are becoming more abundant than others (F_1,79 = 71.85, p < 0.0001, slope = -0.00177±0.000209; Figure 2d.  Evenness did not differ among cover crop treatments (F_3,9 = 1.36, p = 0.3149; Table 1).  Shannon diversity also declined with depth (F_1,79 = 113.36, p < 0.0001; Figure 2f).  These results mirrored those seen in the fall, demonstrating that as I move through the soil profile, measures of diversity (evenness and richness) decline.   

              The first three principal components from analysis of the EcoPlate™ output explained 62.3%  of the variation in the microbial metabolism of the 31 carbon sources.  As in the fall, all functional groups loaded heavily on PC 1 (Table 2; Supplementary Table 1).  PC 2 was dominated by carbohydrates, and PC 3 had all functional groups represented except for amino acids (Table 2).  Analysis of the factors revealed that the cover crop present in the spring significantly affected PC 2 (F_3,9 = 4.16, p = 0.0418).  Cereal rye plots scored high in PC 3 and low in PC 2 (Figure 4b).  This indicates that soil in the cereal rye plots included a greater density or diversity of microbes capable of metabolizing a range of carbohydrates compared to soil from mixed cover crop plots (t_3,9 = -3.04, p = 0.0565).  Cereal rye tended to separate out, while all other treatments and the control clustered together (Figure 4b).  Sampling depth had a significant impact on PC 1 (F_1,79 = 108.43, p < 0.0001; slope = -0.0433±0.0042), PC 2 (F_1,79 = 5.93, p = 0.0171; slope = -0.0146±0.006), and PC 3 (F_1,79 = 5.39, p = 0.0228; slope = -0.0139±0.006) with a decline in metabolism of carbon sources with depth.  As seen in the fall, as depth increased, values for all PCs declines.  

              Usage of all functional groups decreased across depth (Figure 6, Table 4).  Similar to the fall, the trend was that phenolic compounds were least utilized across all depths, but in the spring they were utilized more by cereal rye than by the other treatments and the control group (F_3,9 = 7.78, p = 0.0072; Table 3).  At shallower depths, amino acids and carboxylic acids tended to be greater utilized.  The usage of carbohydrates was also greater at the surface than at all other depths.                

4. Discussion

              The aim of this study was to investigate how the ability of sampled microbes to metabolize 31 types of carbon as a measure of community function changes across depths and among different cover crop treatments using community level physiological profiling with BIOLOG EcoPlates ™.  Because plant root growth adds carbon and affects rhizosphere properties, I postulated that soil depth as well as the presence and type of cover crop would significantly affect the soil microbial community. I found a consistent trend in regard to depth that accorded with predictions whereas trends associated with cover crop treatment were equivocal.  

For all measures of the soil microbial community, microbial activity declined linearly with soil depth and distance from  .  Microbes sampled near the surface metabolized nearly all of the carbon sources.  Variation in the degree of metabolism of these carbon sources was also lower near the surface.  However, even within the top 15 cm of the soil profile that is normally homogenized for soil analysis, the trend in reduced activity with depth was evident in metabolism of all carbon sources. As sampling continued deeper into the soil profile, fewer carbon sources were metabolized and the variation in degree of metabolism increased.  Both the number of carbon sources used and the evenness in use contribute to diversity as reflected in the Shannon diversity index, which declined as I sampled deeper through the soil profile and further outside of the rhizosphere.  This relationship between depth and functional diversity is likely due to the decline in resource availability and reduction in carbon quality that occurs farther outside the rhizosphere (Fierera et al., 2002).  At the surface more crop residues are available for microbes to decompose, and in the top of the soil profile greater plant root growth provides root exudates to nourish the microbial community (Rankoth et al., 2019). 

              As predicted, the functional groups metabolized by the microbial community also changed across depths.  While the overall metabolism of all functional groups decreased across depth, there were also shifts in which groups were most utilized.   Closer to the soil surface, these differences among functional groups metabolized by soil microbes were more apparent.  In the fall, polymers were the most utilized substrate group closer to the soil surface.  Biolog EcoPlates™ contain four polymers:   .  Trends in the impact of the cover crops on the soil microbial community were less clear.  In the fall, the cover crop planted the previous season significantly affected PC 2 and PC 3.  This  indicates that the cover crop planted the previous season has some lasting effects on the microbial community, even after the soy cash crop grew through the following summer.  However in the spring, the cover crop planted only had a significant effect on PC 2.  This is likely due to the poor establishment and delayed germination observed. 

              In the analysis of spring samples, I found the microbial community metabolized fewer  carbon sources and had greater variance in their use, resulting in lower diversity.  The six functional groups were also utilized to different extents in the fall and the spring, with shifts in which groups were most commonly metabolized at each depth increment.  In both the fall and the spring, though, phenolic compounds were the least metabolized.   While samples were collected and processed at different times and are thus not directly comparable, trends can still be noted.  The principal components analysis revealed a different pattern in the trend of carbon source metabolism in the fall and the spring .  EcoPlate™ output explained less variation in the microbial metabolism of the 31 carbon sources in the spring than in the fall (74.8% and 62.3%, respectively).  In the fall, pennycress and control plots tended to separate out, with cereal rye and mixed cover clustered together.  In the spring, though, cereal rye separated out with all other treatments clustered together.  One explanation is that there may be a seasonal component contributing to changes in the microbial community.  Other studies have shown that seasonal variation, such as precipitation and temperature fluctuations, can impact the activity of the soil microbial community (Hamel et al., 2006).  Additionally, the amount of available carbon changes throughout the growing season, with less being available when there is less vegetative cover at the soil surface (Kaiser and Heinemeyer, 1993).

              The trend I observed with depth was consistent across seasons, but the trend for crop was not.  In the fall, I observed more differences in diversity measures between experimental plots, despite all plots having soybeans planted most recently on them.  In the spring, the diversity measures were more consistent across experimental plots, despite significant differences in the amount of plant coverage.   This seems to indicate that better establishment of the cover crop (like in the case of cereal rye) does not necessarily lead to increases in measures of diversity.  Control plots had similar diversity measures to treatment plots.  This seems to indicate that the diversity of live plants may be equally important to successful establishment when it comes to microbial diversity.  Control plots contained a wide variety of weeds present throughout the spring, and often had vegetation present before the germination of some cover crops (mixed cover treatment).  This may indicate that with better establishment, a mixed cover crop treatment can positively impact soil microbial diversity.  The results that I observed in the spring are likely due to a combination of factors: soil moisture was much lower in the spring than in the fall, and plant coverage was lower and more variable in the spring than in the fall.  Cover crop establishment was incredibly variable in the spring, with cereal rye having the most successful establishment and the mixed cover crop having the least success. 

              Previous studies have shown EcoPlates™ to be an effective method for examining changes in the soil microbial community in agroecosystems (Bonanomi et al., 2020; Gomez et al., 2006; Zhou et al., 2008).  EcoPlate™ studies can provide support for further work using molecular techniques that can provide more detailed information about the structure of the microbial community.  Microbial biomass and bacteria:fungi ratios provided by molecular techniques, like PLFA and EL-FAME, can be useful at determining the impact of cover crops on the microbial community.  As a culture method, EcoPlates™  cannot capture the full extent of what is happening in the soil.  Cover crops may be impacting total microbial biomass, but not the overall functional diversity or species diversity of the microbial community (Frasier et al., 2016).  Additionally, EcoPlates™ only contain 31 carbon sources, which may not capture all those able to be metabolized by a microbial community.