REPORT FROM 2018

Our first task in early spring 2018 was to determine our final list of groundcover species for our research.  In our original proposal we had selected eight perennial species with which to design four distinct polycultural treatments, in addition to a treatment using only Dutch white clover (Trifolium repens) and a control treatment of woodchips only, for a total of six different treatments.  After working with our consultant Dave Jacke, we refined and simplified our experimental treatments to address one of three distinct objectives: 1) Soil/mineral building 2) Beneficial insect/pollinators and 3) Nitrogen-fixers, for a grand total of four treatments including a woodchip control.  We briefly considered creating both a "conventional" and "alternative" nitrogen-fixing polyculture but ultimately were unable to identify enough woody perennial n-fixers that would qualify as groundcovers.   We elected to direct seed a variety of herbaceous perennial clovers and lupines for our n-fixing groundcover treatment instead. 

To design the two primary soil building and beneficial insect polycultures we narrowed down a much larger plant list using the following criteria:

1. Not a sterile cultivar or overly hybridized for use in landscaping
2. Relatively available and easy to acquire from a nursery at an affordable price
3. Must be hardy to zone 5 at least.
4. Anything planted from seed could not need more than 60 days of cold stratification (we only had 3 months before the warm growing season began)
5. Preference for edible plants over similar functioning plants that were not edible i.e. peppermint vs. ajuga mint

Each polyculture also needed a mix of types of groundcovers such as creeping, spreading, clumping, etc., as well as a mix of tolerance to sun and shade, wet and dry conditions.

Our final list was as follows:

We were ultimately successful in sourcing everything we selected.  To save money we seeded the mints and aster in flats on our farm as they are fairly easy to grow, and we sourced oregano and wild violets from neighbors.  We also sourced comfrey divisions from stock we already had growing.  All the remaining plants were sourced from a small local nursery, with the exception of three less common species (C. verticillate, C. plumbagenoides, and G. macrorrhizum) which were obtained from a larger nursery.

After designing our polycultural treatments and sourcing our stock, we worked with our Technical Advisor Liz Garofalo on the experimental design of our field layout.  Because our orchard is planted on contour, each row is somewhat curved and varies in length from 170' to 440'.  This year's hazelnuts were planted in newly dug swale/berm rows #14 and #15 which measured 180' and 420' respectively (for a total of 600 row feet).  Originally we had proposed six different treatments in 100' rows (each encompassing 33 hazelnuts planted 3' on center).

Instead of planting 100' long rows of each treatment, we redesigned our experimental layout to incorporate treatment repetitions and thus minimize the effect of variables related to existing conditions in our field (i.e. differences in soils, sun exposure, moisture, etc.). 

With only four different treatments in our experiment (Soi/mineral builder, Beneficial/pollinators, N-fixing, and control) we shorted each treatment to a 30' row (encompassing 10 hazelnuts planted 3' on center) and randomly repeated each treatment a total of five times over the total 600 available row feet.

Our final field layout is as follows:

North                                                                                                                                       South

Row 14 (180')

Row 15 (420')

S=Soil/mineral builder

P=Pollinator/beneficial insect

N=Nitrogen fixing

C=Control

The layout of the individual plants within the two primary polycultures ("S" and "P") came out of a few basic considerations: the availability of moisture across the berm, with more water near the base and drier conditions at the top; sun exposure, with more shade developing around and between the hazelnuts over time; and the habits of the groundcovers themselves, as we tried to minimize competition for horizontal and vertical space while still attaining good overall coverage for weed control.  We elected to plant a "pattern within a pattern," as in our original proposal by repeating an identical 3' block of plants 10 times over the length of each of the five 30' treatments.

The layout of the primary polycultures is as follows (bird's eye view):

Soil/mineral builder ("S"):

Beneficial insect/pollinator ("P"):

We planted 200 hazelnuts as bare root stock on May 15th-16th, 2018 into freshly dug berms.  Prior to planting, we took a single representative soil sample of both rows, and then covered all bare soil with a 2" layer of fresh ramial woodchips.  We also soaked overnight and broadcast seeded the nitrogen-fixing species into the N treatment plots and covered with a partially-decomposed layer of woodchip mulch.  Later that week we started three flats of catmint, fieldmint and aster seed. Unfortunately the field mint germinated poorly and we were unable to include it in our Soil/mineral building polyculture as we did not have enough plants to distribute.

The following week we received and planted plugs of mountain mint, yarrow, chives, and oregano.  At the end of May we received and planted the geranium and coreopsis.  Unfortunately the plumbago was backordered and would be not be shipped until mid-July.  In early June we obtained and planted wild violets and comfrey.  In mid-July we planted out our flats of catmint and aster, along with the plumbago. We also decided to reseed our nitrogen-fixing treatments as germination had been spotty due to a dryer than normal June.  Germination was much improved with good rainfall in July.  The perennial lupines also did not germinate well when we planted them in May, but unfortunately we did not have any leftover seed to replant.

At the end of July we met again with Liz to determine our precise measurement methods and data collection techniques.  We designed a field note template to record:

1) Diameter of primary hazelnut cane at 4" up from ground - to the nearest 1/32 (using calipers)

2) Primary and secondary shoot growth on hazelnuts

3) Number of suckers growing at the base of each hazelnut

4) Goundcover coverage on a scale of 0-5 (0 being unchanged; 5 being complete coverage) on either side of each hazelnut with the tree in the center (a 3' span, 18" on each side).  For the control plots, N/A was recorded for coverage.

We collected this data on the middle four hazelnuts in each 30' treatment (#5, #6, #7, #8) to minimize crossover effects of adjacent treatment plots.  We also noted other additional observations such as browse or dieback.  We opted not at this time to collect hazelnut leaves for a tissue analysis; this will be included in next year's data collection.

Lastly with Liz's help we set out three large yellow sticky cards on stakes in our trial area to collect a sampling of insect specimen.  We left them up for a period of five days and then collected and replaced with new cards for another period of five days.  Unfortunately during both periods of insect collection we had an abundance of stormy weather and rain which soaked the cards and made it more difficult to precisely ID the insects.  We hope to still have enough baseline information to assess changes in insect populations in future years.

At certain intervals throughout the growing season we hand weeded the entire trial area, a task which did become less onerous as the groundcovers grew.  The primary weeds we were managing were: wild lettuce, brambles (blackberry, raspberry, dewberry), pokeweed, goldenrod, meadowsweet, buckthorn, bindweed, stump sprouts (birch, maple, oak, cherry), and ferns.  In total we weeded five times on June 25, July 9, July 16, Aug 6, and Sept 30 for a total of 20 person-hours.

In early October 2018 we again collected measurement data on the hazelnuts using the same methods.  A fall soil sample was also collected on Oct. 21st, this time taking two samples at a depth of 6" from the middle of each plot and mixing together the soils from the same type of treatment, for a total of four soil samples (S, P, N, C).

REPORT FROM 2019

In the second year of our study, we focused on observing the interspecific traits of the groundcovers and hazels with a less managed approach.  In late-spring we observed vigorous regrowth of the following species:

These groundcovers exhibited favorable traits such as dense clumping and/or vigorous spreading beneath the hazelnuts. 

In June we spent about 12 person-hours hand weeding the trial plots  of any brambles, wild lettuce, meadowsweet, black locust, and buckthorn emerging in the rows.  The rest of the season we spent just 20 person-hours on weeding - the rest of our time and manual labor was spent on data collection and mowing the alleys between the rows as part of a regular field maintenance schedule.

At the end of June we took our first measurements to compare summer hazelnut growth and groundcover spread.  Again we recorded information on four points:

1) Diameter of primary hazelnut cane at 4" up from ground - to the nearest 1/32 (using calipers)

2) Primary and secondary shoot growth on hazelnuts

3) Number of suckers growing at the base of each hazelnut

4) Goundcover coverage on a scale of 0-5 (0 being unchanged; 5 being complete coverage) on either side of each hazelnut with the tree in the center (a 3' span, 18" on each side).  For the control plots, N/A was recorded for coverage.

By mid-summer, we observed that some of the perennial groundcover species had not performed well in their second year.  The following species were either difficult to locate or missing altogether:

In mid-August we did our first leaf tissue sample analysis.  We used a method similar to taking soil samples to randomly collect ten hazelnut leaves from the middle four hazelnut plants in each trial plot and combined the leaves receiving the same groundcover treatment into separate bags.  This resulted in a total sample size of fifty leaves from each of the four treatment types, which was sent to the UMass Leaf Tissue Lab.

In mid-October we took our second round of soil samples using the same method as the previous fall.  The soil samples were sent to Logan Labs, LLC.  The following week we took our second round of measurements of the hazelnuts and groundcover spread in the second year, with one difference: because the hazelnuts had mostly lost their foliage, identifying and counting suckers became a difficult and time consuming task.  Since we had already collected this data in June we opted not to include this number in our data set for October, assuming that no new suckers would have grown since June and any dieback (or new growth) of suckers would be captured in our data next season.

Unfortunately we were unsuccessful in collecting data on the yellow sticky cards in our second year on the presence of pollinator and insect diversity in the trial plots.  However, anecdotal observation supported an increased presence of beneficial insects due to the abundant supply of flowering species, starting with the geranium, followed closely by the chives, and continuing for the rest of the season with the yarrow and oregano.  The comfrey and mountain mint also flowered abundantly for several weeks mid-season.

In our experimental design we also intended to re-woodchip the control beds; this did not occur as planned, and some soil erosion was noted in areas where the original woodchip layer had already broken down.  A new layer of woodchip mulch will be re-applied in the spring of 2020.

Lastly, at the end of October, we offered a field trip opportunity to the Smith College students in Professor James Grogan's Biological Economics class in their module on fruit and nut trees.  During the field trip we discussed our SARE research and the observations we have made so far on the interspecific relationships between the hazelnut and the perennial groundcover polycultures.

REPORT FROM 2020

In the third year of our study, we continued to observe the interspecific traits of the groundcovers and hazels with a less managed approach.  In late-spring we observed healthy regrowth of the following species:

These groundcovers exhibited favorable traits such as dense clumping and/or vigorous spreading beneath the hazelnuts.  In addition, we saw a decent recovery of white and red clovers, and a patchy distribution of perennial lupins returning in the N-fixing trial plot without any additional reseeding.  In general, we have noted that the trial plots on the southern end of both berms is more lush than the northern ends, where there is more solar gain and thus drier soils.  Since we do not have irrigation in the field, rainfall has played an important role in our experiment.  Soil hydrology is something we did not really take into account and is a important factor to consider in selecting and establishing a successful groundcover polyculture.

In 2019, we spent a total of 32 person-hours hand weeding the trial plots of all unwanted brambles, wild lettuce, meadowsweet, black locust and buckthorn.  This year, 2020, we cut back even more to approximately 16 person-hours.  This was largely due to both the success of previous efforts to remove perennial weeds and the increasing coverage and root mass of competitive groundcovers: comfrey, oregano, and yarrow especially.

At the end of June we took our first measurements to compare summer hazelnut growth and groundcover spread.  In previous years, when the hazelnuts were young whips, it was easy to find and measure the "primary" and "secondary" shoots from stem to tip to get a sense of that year's foliar vigor.  This year, many of the hazelnuts began to develop more bush-like characteristics, with some suckers exceeding new growth from the main stem.  As a result, we opted not to make guesses about which branches to measure and instead collected data on the overall "plant width" from a birds eye perspective.  We regret not taking this measurement from the beginning of the study as it would have been a more insightful number to compare in drawing conclusions.  We repeated these same measurements in September.

The four measurements we recorded were:

1) Diameter of primary hazelnut cane at 4" up from ground - to the nearest 1/32 (using calipers)

2) Total plant width, measured from a bird's eye perspective

3) Number of suckers growing at the base of each hazelnut (June only*)

4) Groundcover coverage on a scale of 0-5 (0 being unchanged; 5 being complete coverage) on either side of each hazelnut with the tree in the center (a 3' span, 18" on each side).  For the control plots, N/A was recorded for coverage.

*As in the previous year, we opted not to count suckers in the fall as it was difficult to do accurately and was unlikely to have changed significantly since the count in June.

In mid-August we prepared leaf tissue samples, collecting ten leaves from the middle four hazelnut plants in each trial plot and combining all the ones from the same groundcover treatment.  Unfortunately, we were unable to get the samples analyzed at the UMass Leaf Tissue Lab due to closures in response to COVID-19.  We attempted to find another lab to send them to, but could not confirm exactly which test would be comparable to the one we had used last year and could not reach any technicians for support.

For this reason, we have decided to extend our study for an additional season and plan to collect all measurements and data for analysis at the end of the fourth year (2021).

In mid-October we took our third round of soil samples and sent them to Logan Labs, LLC.  While we do not yet have statistically significant data to report between the trial plots, it is encouraging to see overall improvements in soil health, pH, CEC, and organic matter.  We mostly attribute this to allowing the disturbed soils to resettle in no-till berms as perennial roots penetrate and stabilize long term plant-soil interactions, but look forward to seeing if there is any correlation between plot treatments and soil nutrient availability.

This year we made a renewed effort to collect insect population data using yellow sticky cards in the field as we did in 2018.  We are still seeking support to assist with insect ID to determine if there is any empirical data showing an increase in pollinator and beneficial insects.  Again, anecdotally, we observed low rates of predation from leaf eaters and did not see any tent moths or Japanese beetles in the field, possibly due to climactic factors, but still in support of  healthy looking plants.

We also made sure to repair areas that had suffered damage from erosion and wildlife (most likely deer trampling), and gave the control plots a fresh layer of woodchips after a thorough weeding in early summer.  The ramial chips we used were leftover from prior years and thus already partially-decomposed. 

Lastly, we completed our planned outreach activity with a workshop presentation at the Northeast Organic Farming Association (NOFA Mass) Winter Conference.  Our virtual workshop was titled, "Growing Nuts: Land Management and Soil Building," and featured an overview of our establishment techniques and low-carbon maintenance of our perennial tree crops, and included a thorough report of our SARE-funded research to date.  The workshop lasted 90 minutes and engaged around 40 virtual attendees.  A recording of the workshop is available through NOFA Mass.

We are still considering offering a field day opportunity in Summer 2021 (pandemic-permitting) and have tentatively reached out to our mentor, Dave Jacke, about co-presenting a day-long workshop here on the farm.  We also hope to submit our final report and results for publication in the 'Nutshell,' the quarterly publication of the Northern Nut Growers Association.