Bryan Markhart Ecological Theory and Research
Return to Advanced Ecology 2012
Return to Bryan Markhart
Weed science research focuses on the characteristics of weed communities in agricultural fields with the applied relevance of creating sound weed management strategies. Good weed management practices are tailored to specific fields with goal of addressing both short term and long-term weed pressure. To accomplish this weed management must not only remove weed pressure in the on crops during the immediate season, but also work to degrade the soil seed bank, thereby reducing the number of seeds present in the soil and preventing serious weed problems in the future. No seeds means no weeds, which helps boost yields, but also dramatically reduces the mount of input energy required, whether it be in the form of tractor fuel or labor. This saves time and money for the farmer. Therefore weed management strategies should be primarily aimed at addressing long-term seed banks.
Agricultural progressives in the latter half of the 20th century have been championing a form of weed control referred to as integrated weed management (IWM). This strategy relies on various different forms of weed control responding to specific issues in fields (Cussans, 1993). With proper use IWM techniques some farms are able to all but completely eliminate the soil seed bank (Nordell, 2009).
In most conventional agricultural practices today the primary weed management tool is tillage, turning the soil and thereby killing, or at least significantly stunting weeds. However, one must ask how effective this conventional practice is at ebbing future weed pressure. This has been the primary focus of this study: to test the long-term effects of conventional tillage practices (moldboard plough followed by disking) on soil seed bank characteristics. As a result of the circumstances available, I looked specifically at the affect of disturbance due to 3-4 years of continuous conventional corn cultivation on soil seed bank abundance and community composition. This is essentially a study looking at succession patterns in response to disturbance in soil seed banks. This was all with purpose of discovering the effectiveness as a weed management.
Hypothesis: I predicted that there would be significant differences in seed bank abundance and community composition between fields at different successional stages after abandonment from corn production. I also predicted an observed decreased in seed abundance in more recently cultivated fields suggesting that conventional tillage practices employed are an effective at long term weed management practice. The null hypothesis being that there is no significant correlation between seed bank abundance and time since cultivation suggesting conventional tillage is an ineffective form of long term weed management.
For this experiment I used the circumstances available to me on Bennington College campus. For many years a local dairy farmer has used college land to grow feed corn for his herd. He will use each field in continuous corn production with management beyond tilling, seeding and harvesting. Every four to five years he will rotate the field in cultivation allowing the abandoned field to follow natural succession with no further soil disturbance. Each field has been hayed annually. Some of the land use history records used in this study are not precise and there may be some gabs, but this information was the best available. I used the following methods to collect my seed banks data.
Plot Selection: I examined the seed banks in Long Meadow (Long), the field below the orchard (Orch), Jennings Meadow (Jenn), and the field beyond the end the world (End) (Map 1). Long as the field cultivated for the past three years, Orch was cultivated before Long and Jenn before that. End was chosen as a long-time-since-cultivation field as the soil has not be disturbed in almost 60 years. I collected soil to the depth of the farmer’s tillage, based on the assumption that the tillage techniques fairly thoroughly mix the soil in this horizon, evenly distributing the seeds after tillage.
Sample collection: I chose three samples plots per field to collect soil from and took two 15cm deep cores using a volumetric soil sampler to at each plot. The collection plots were roughly evenly distributed throughout each field by imagining a line bisecting each field across their longest axis and evenly spacing the collection cites along this line. I then combined the samples from each plot, breaking up clumps and removing plant matter and spread one brimming 500 mL beakers of soil from each plot into flats. The flats were first laid with a layer of reemay, followed by two brimming 500 mL beakers of potting soil to retain moisture and provide a growing medium. The soil was laid about 2-3mm think. Growing method adapted from Ter Herrdt et al. (1996).
Seed bank measurement: Soil samples were allowed to grow for 28 days under less than ideal, but not terrible growing conditions: room temperature; water every 2-4 days, and approximately 6.5hrs sun daily, as available. The soils samples experienced several slight droughts during this time, which may have cause, some seedling mortality. Seedlings were counted on day 28. Type and abundance were recorded. Seedlings were not identified according to taxonomic rules, but simply as like-taxa instead code names were assigned
Seed bank analysis: In order to get a picture of seed bank composition in each field, seed abundance, richness, and Shannon-Wiener index values were calculated at plot and field levels. To determine the statistical significance of my data I used these calculations to perform ANOVA and pairwise comparison tests on abundance and richness data. To compare the relative variations in community composition I created a Principal Components Analysis (PCA) ordination.
I counted a total of 18 distinct taxa (Table 1). Non-descript seedlings were classified as having small round cotyledons (SRC), or long cotyledons (LC).
There were clearly differences in the abundance and diversity at the plot scale, however the overall importance of this at the field level was statistically insignificant except for Orch. There was no distinct trend in the abundance, richness, or Shannon-Wienner diversity index in relation to the fields’ time since cultivation.
The results of the ANOVA suggests significance in the degree of variation in seed abundance between the four fields, p=0.06. The pairewise comparison results suggests the possibility of significant differences in abundance at field level between Orch and both Jenn and End: p=0.06, p=0.09, respectively (Figure 1), as well as for the differences in richness between these fields: p=0.09, 0.06 respectively (Figure 2). There were no statistical differences in the Shannon-Weinner indexes between any of the fields.
The PCA ordination shows clear differences in the community composition of each field, most obviously the orchard field is distinctly different from the other three, however it is not uniformly dissimilar from the other three (Figure 3). This difference is the only statistically significant one, however each field does seem to have a definite identity of its own (word choice?).
While my results do not show strong statistical significance or any definite trends in the relationship between a field’s time since cultivation and either seed bank abundance or community composition, there still is a possible trend suggested here. The slight significance in the difference in abundance, richness, and community composition of Orch compared to the other fields suggests that a relationship between seed bank characteristics and tillage may exist, however, in order to draw any definite or clear conclusions further research must be conducted.
The results I obtained in this experiment supported my null-hypothesis, that conventional tillage due to corn production would have no affect on the long-term soil seed bank of a field. Some relationship may exist, but no conclusions can be drawn at this point However the relationship suggested by the higher abundance and richness observed in Orch rejects my hypothesis that tillage should reduce soil seed bank abundance. Further research and a larger data set is required to draw definite conclusions, but the slight statistical significance in Orch suggests that there may be a variable creating distinct seed banks between these fields. My current results do not support conventional tillage as an affective weed management technique, as it showed no affect on the soil seed bank. However it is also possible that methodological issues with regular watering of the samples may have affected the outcome of the seed bank analysis.
Future research should further explore the possible relationship determined in this preliminary experiment. In order to obtain results that provide a more complete picture of the soil seed bank the experiment should include a much larger sample size looking at more fields with more timescale resolution and using more samples plots per field. During the growing process, the experimenter should provide a thicker base layer of potting soil in the trays to prevent them from drying out as quickly. They should also take weekly counts of seeds and then remove counted seedlings; this does not require every germinated seed to survive the entire growing period and will buffer against death by drought. I would also be interesting to examine the affect that tillage has on seed dispersal through the soil horizon.
Super special thank you to Kerry Woods, David Norman, and the rest of the ecology class foe helping me work through the project. Also, research assistants Zofia Lamprecht and Zoe Banfield for help with soil collection.
Cardina, J., Regnier, E., & Harrison, K. (1991). Long-term tillage effects on seed banks in three ohio soils. Weed Science, 39(2), 189-1994. <http://www.jstor.org/discover/10.2307/2390273?uid=3739952&uid=2129&uid=2&uid=70&uid=4&uid=3739256&sid=56238940043>
Ter Herrdt, G. N. J., Verweij, G. L., Bekker, R. M., & Bakker, J. P. (1996). An improved method for seed-bank analysis: Seedling emergence after removing the soil by sieving. Functional Ecology, 10(1), 144-151. <http://www.jstor.org/stable/10.2307/2390273>