Understanding production and conservation tradeoffs of vertical tillage practices

We conducted on-farm field trials to compare no-till and vertical tillage treatments on nine farms across Lancaster County (n=6) and Chester County (n=3) in southeastern Pennsylvania in 2021 and replicated these on-farm field trials on 12 farms across Lancaster County (n=10) and Chester County (n=2) in 2022.  These cooperating farms consisted mostly of cash grain operations raising corn (Zea mays L.), soybeans (Glycine max L.), winter wheat (Triticum aestivum L.), and winter barley (Hordeum vulgare L.).  Well-drained silt loam soils coupled with frequent manure applications (and occasional applications of spent mushroom substrate) contribute to relatively fertile and historically high-yielding environments on all cooperating farms.  Many of the farms utilize no-till and cover cropping practices.

Experimental Design

On-farm strip trials comparing no-till and vertical tillage were conducted using a nested treatment structure due to variability in tillage legacy and equipment type across farms.  We used tillage management legacy as a grouping factor with two levels: 1) fields where long-term no-till has been practiced for approximately 10 years or more and where vertical tillage was introduced (n=10); and 2) fields where vertical tillage has occurred annually for the previous six to eight years (n=10).  In total, ten tillage treatment comparisons were conducted at locations with each tillage management legacy (with or without a history of vertical tillage), resulting in 20 pairwise tillage treatment comparisons and 40 experimental units.  At each cooperating farm, strip trials were placed in either multiple fields or unique locations within large acreage fields that contain variability in soil conditions and/or terrain. These trials occurred in fields being rotated from corn grain to full-season soybeans.  In these fields, the corn stubble remained unharvested and undisturbed over the winter.  Tillage treatments were employed in the spring prior to planting in paired and randomly located field-length strips that included a no-till (NT) control strip and a vertical tillage (VT) strip.

Experimental Design

Field Operations

Farmer cooperators employed vertical tillage using owned or rented implements at an average working depth of 5 cm to manage corn residue in the early spring (April) prior to establishing full-season soybeans.  The vertical tillage tools were compliant with standards for “Low-Disturbance Residue Management Equipment” as defined by the Pennsylvania Resource Enhancement and Protection (REAP) Program Guidelines for fiscal year 2019 (REAP Program Guidelines, 2019).  These guidelines stipulated that qualifying equipment should be set as follows: (1) disc blade angle must not exceed five degrees, (2) disc blades must have no concavity, (3) working depth of equipment must not exceed four inches and (4) minimum surface residue cover must not fall below 60% throughout the year (REAP Program Guidelines, 2019).  Three types of common vertical tillage implements used on cooperating farms included a Salford Independent series tool (Salford Group, Inc., Salford, ON, Canada), a Great Plains Turbo-Till (Great Plains Manufacturing, Inc., Salina, KS), and a Kuhn-Krause Excelerator (Kuhn North America, Inc., Brodhead, WI).

Each cooperating farm implemented standard fertility, seed protection, and crop protection programs for full-season soybeans.  A burndown and pre-emerge herbicide program were implemented on each farm and herbicide product selection was determined by the cooperator or their custom pesticide applicator.  Additionally, all cooperators applied a post-emerge soybean herbicide product(s).  Harvest data was collected using either a yield monitor or by using a weigh wagon or truck scale, and a moisture tester.  Strip width was based on the size of available harvesting equipment so one or two combine passes could be completed within tillage treatment strips.

Data Collection and Data Analysis

1. Soil conservation metric:  The proportion of surface residue cover was determined after vertical tillage treatments are implemented (April) using the line-transect method following the USDA-NRCS standard surface residue cover assessment protocol (Estimating crop residue cover, 1984).

2. Weed management metrics:  We measured winter annual weed abundance in the early spring (April) after tillage treatments were employed and just prior to soybean planting to evaluate the potential for vertical tillage to be employed as a pre-plant weed control tool.  We also measured summer annual weed abundance at the early spring (April) sampling point to assess vertical tillage effects on recruitment of early-emerging summer annual weed species, which may facilitate the use of false-seedbed tactics in no-till systems.  To assess weed abundance, we utilized a belt-transect sampling approach.  Three 15 m transects were established in unique field positions (toe slope, foot slope, backslope) within each strip and co-located between paired strips (i.e., 3 paired transects; 6 transects per paired strip location).  The location of these three paired transects were marked using geo-referencing software (i.e., QGIS) and these locations were used for data collection later in the trial.  Within each transect, the presence or absence of weeds located within 15 cm on each side of the transect was recorded every 15 cm lengthwise.  Presence data was expressed as a proportion of the total number of observations per transect (n=100) as an estimate of weed abundance.  We assessed weed abundance in early summer (June) prior to a POST herbicide application using the same sample methodology to assess recruitment patterns of summer annual weeds and residual herbicide efficacy between no-till and vertical tillage strips.

3. Soybean performance and slug feeding metrics:  Soybean emergence was assessed by measuring the emerged plant population and soybean growth stages when soybean plants reached the Cotyledons Expanded (VC) growth stage.  A 5.3 m long transect (equivalent to 17.5 ft or 1/1,000^th of an acre) was established in each strip at each previously georeferenced location. Emerged soybean plants were counted along transects and soybean stand establishment was expressed on a plants per square meter basis.  Soybean stand uniformity was assessed at the time of population assessments using the same transect.  Twenty plants were counted along the transect, their growth stages recorded, and an average growth stage determined.  The growth stages determined at each of the three data collection waypoints within each strip were averaged.  Crop yield was measured using a yield monitor on the combine or by measuring crop mass harvested from each strip using a weigh wagon or truck scale, and a moisture tester.  By measuring crop yield at harvest, we evaluated a tangible and direct result of changing a field management practice.

The incidence of slug damage was assessed in late spring at the V1 to V3 soybean growth stage by establishing two separate transects at each data collection waypoint within each strip.  Within each transect, the number of soybean plants per three meters of row were counted and assigned a slug damage severity rating based on leaf defoliation (%) via slug feeding where 0 = no apparent damage, 1 = 0-25% defoliation, 2 = 26-50% defoliation, 3 = 51-75% defoliation, and 4 = 76-100% defoliation (Douglas & Tooker, 2012).  The number of total plants damaged per three meters of row, and the number of plants damaged based on the ordinal scale, were averaged across the six transects assessed within each strip and expressed as a percentage.

4. Nutrient management metrics:  Soil samples were taken in late fall (October) after soybean harvest was complete to assess nutrient management and soil health metrics (see below).  Within each strip, three soil samples were randomly collected around each georeferenced field position (n =3) used previously for the surface residue cover, weed abundance, and crop emergence metrics.  Soil cores were stratified by sampling depth (0-2.5, 2.5-7.5, 7.5-15 cm) and then composited.  Basic soil fertility analyses was conducted at AASL, including soil pH using 1:1 water extracts (Eckert & Sims, 1982), cation exchange capacity (CEC) using summation of cations (Ross & Ketterings, 2011), base saturation using calculation of cations (Sikora & Moore-Kucera, 2014), extractable phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), sulfur (S), zinc (Zn), and copper (Cu) using a Mehlich-3 (ICP) soil extractant (Wolf & Beegle, 2011), and soil organic matter using loss on ignition (LOI) (Schulte, 2011).  Using these soil test results, both pH and soil test phosphorus stratification were assessed.

5. Soil health metrics:  Permanganate oxidizable carbon (POXC) was measured using methods described in Weil et al. (2003), where the active carbon fraction of soil organic matter was oxidized with a weak solution of potassium permanganate (KMnO₄).  The assessment was completed by adding 20 mL of 0.02 M potassium permanganate solution to 2.5 g of air-dry soil placed in a plastic centrifuge tube.  The centrifuge tubes were shaken on a mechanical shaker for two minutes at 120 strokes per minute and then placed in a holding rack for ten minutes while soil particles flocculate and settle on the bottom of the tube.  A pocket colorimeter (Pocket Colorimeter II, Hach Company, Loveland, CO) was used to read the absorbance value for the color concentration of the solution with POXC.  The lighter the purple color of the solution with POXC, the less absorbance measured by the colorimeter, the more active carbon in the soil sample which was oxidized.  POXC is expressed as mg C kg^-1 soil for each sample.

Microbial respiration was analyzed using the CO₂-burst method described in Franzluebbers et al. (2016) to measure the flush of carbon dioxide (CO₂) from a sample of air-dry soil re-wetted to 50% water-filled pore space after a 24-hr aerobic incubation in a dark setting at room temperature.  To complete the incubation, 10 g of air-dry soil sieved to 2 mm was placed in a 50 mL plastic beaker and distilled water was added to the top of the beaker to reach 50% water-filled pore space as determined by a previously measured sample specific bulk density.  The plastic beaker was then be placed in a 473 mL (1 pint) glass canning jar sealed with a screw band lid (Ball Corporation, Broomfield, CO) that has a rubber septum installed within the lid and sealed with vacuum grease.  After 24 hours, a syringe was used to extract a 1 mL sample of headspace air from the jar and this sample was injected into an infrared gas analyzer (LI-7000 CO₂/H₂O Gas Analyzer, LI-COR Biosciences, Lincoln, NE) where the concentration of CO₂ in the sample air was compared to that of a reference air with a known CO₂ concentration.  The concentration of CO₂ in the headspace of the jar was converted to mg C while accounting for the concentration of CO₂ in the ambient air in the room when the jars were sealed at the beginning of the incubation.  Microbial respiration estimates were expressed as mg CO₂-C kg^-1 soil (1 d)^-1 for each soil sample.

Wet aggregate stability was assessed utilizing methods described in Kemper et al. (1986), whereby an air-dry soil sample consisting of 1-2 mm sized aggregates was subjected to the disturbance from a wet sieving apparatus and an ultrasonic probe to determine the difference between water stable and water unstable soil aggregates.  For wet aggregate stability analysis, 4 g of air-dry soil 1-2 mm in size were placed in a 0.5 mm sieve and subsequently placed in a wet-sieving apparatus (Five Star Scientific, Twin Falls, ID) and into a pre-weighed metal can filled with 75% distilled water.  The wet-sieving apparatus was mechanically raised and lowered 1.3 cm at a rate of about 35 times per minute, as eight soil samples, each in their respective sieves and cans, were raised and lowered in water for three minutes.  Each soil sample in its respective sieve was then removed from the first set of cans and placed in a second set of pre-weighed metal cans filled partially with distilled water.  These sieves, now inside a second set of metal cans, were placed under an ultrasonic probe as part of a sonifier (Sonifier Cell Disruptor Model W185, Heat-Systems-Ultrasonics, Inc., Plainview, L.I., NY) for approximately 30 seconds that dispersed any remaining soil aggregates into primary particles.  Both sets of cans were placed in a 110°C drying oven overnight to evaporate all water from the cans.  The mass of both sets of cans was recorded and the mass of the dry soil particles from the first set of cans, representing the portion of water unstable aggregates, and the mass of the dry soil particles from the second set of cans, representing the portion of water stable aggregates, was determined.  The percent water stable aggregates (% WSA) in each soil sample was calculated using the following equation:

% WSA = 100 x [Oven dry weight of can B with soil – empty weight of can B] / [(Oven dry weight of can B with soil – empty weight of can B) + (Oven dry weight of can A with soil – empty weight of can A)].

Soil penetration (cone) resistance was also measured at the time of soil sampling in the fall after harvest.  A digital recording penetrometer (Field Scout SC 900 Digital Soil Compaction Meter, Spectrum Technologies, Inc., Aurora, IL) was used to measure soil penetration resistance at three locations around each data collection waypoint within each strip (i.e., nine sampling points total per strip).  Soil penetration resistance in pound-force per square inch (psi) was measured to a depth of 20 cm at 2.5 cm depth increments:  2.5, 5.0, 7.5, 10, 12.5, 15, 17.5, and 20 cm.

Statistical Analysis:  For each response variable described above, we estimated the mean effect size of vertical tillage using linear mixed models by considering tillage treatment, tillage legacy and their interaction as a fixed effect and year, farm (nested within year), and field (nested within farm and year) as random effects.  Mean effect size for each production and conservation metric was reported together to visualize management tradeoffs using our multi-criteria assessment framework.