Does More Diverse Plant Architecture in Pollinator Habitats Influence Native Pollinator and Beneficial Insect Abundance and Diversity?

This study explores whether designing pollinator habitats to include plant species with divergent architecture will increase pollinator and beneficial insect diversity and abundance when compared to habitats containing plants with similar architecture. Plant architecture is the three-dimensional organization of the plant, including shape, size and position of branches, leaves and flowers. Different plant species can have similar architecture. A habitat diverse in terms of species can be similar in terms of architecture. The experimental design accounts for both species’ diversity and plant architecture. It will be conducted on a small organic farm near Colfax, WA owned by Jane Fluegel. The farm is surrounded by large conventional wheat farms. The area selected for the study is a very uniform, flat fertile field. Land will be prepared by tilling, measuring out plots, and laying down paper mulch and sawdust for weed suppression.

Objectives 1) Evaluate if a more diverse plant architecture effects pollinator and beneficial insect abundance and diversity. To accomplish objectives #1:

• We will use a randomized block experiment with five blocks and ten treatments.
• Treatments are designed to account for both plant diversity and plant architecture. The four selected plants are two pairs of non-invasive, commercially obtainable, pollinator friendly plants for the Northwest. Matched pairs are architecturally similar. The two different pairs are architecturally diverse.

Matched plant pairs:

• Salvia Caradonna (S) with Nepata x faassenii ‘Walker’s Low’ catmint (C)
• Leucanthemum x superbum ‘Alaska’ (L) with Gaillardia ‘Burgundy’ (G)

Diverse plant pairs:

• Salvia Caradonna (S) with Leucanthemum x superbum ‘Alaska (L)
• Salvia Caradonna (S) with Gaillardia ‘Burgundy’ (G)
• Nepata x faassenii ‘Walker’s Low’ catmint (C) with Leucanthemum x superbum ‘Alaska (L)
• Nepata x faassenii ‘Walker’s Low’ catmint (C) with Gaillardia ‘Burgundy’ (G)

In addition, each plant will be evaluated separately.

• Each treatment will have 5 replications.
• Blocks control for time of insect counts. We can record insect visitors in ten plots at once.
• Each plant species will be evaluated 1) separately, 2) with an architecturally similarly paired plant species, and 3) with an architecturally dissimilar species (from the opposing pair). This design will allow us to evaluate whether changes in pollinator diversity and abundance are the result of the individual plant (one species in plot), species diversity (two species in plot with very similar architecture) and/or architectural diversity (two species in plot with widely divergent architecture). The ten treatments will be: SS, CC, LL, GG, SC, LG, SG, LC, SL, and CG.
• Treatments will be a 3 x 2 ft. plot with 6 plants randomly planted 12 inches apart.
• Plots will be covered with paper mulch and bark mulch to allow establishment without weed competition. Growers have enough to do without weeding pollinator strips. Full mulch will be applied in spring to reduce winter kill due to crown rot.
• Pollinator visitation and diversity will be assessed at least six full days for each treatment spaced out from April through August. We will use time lapse video camera surveillance. Edwards et al. (2015) demonstrated the effectiveness of portable time lapse video for pollinator counts.
• Time lapse photography will be scanned and pollinator and beneficial insects will be counted and identified. To conserve pollinators, pollinating insects will be identified in a non-interfering manner through photos. Periodic scouting will allow us to become familiar with regular pollinator visitors and to capture unknown insects for positive identification.
• Plant vigor and growth will be accessed each month during the growing season through foliage area cover and flower numbers. Due to the extreme vigor of the plants, this was done by recording percentage plot coverage and bloom coverage of each plant variety. 
• Treatments will be analyzed for significance using ANOVA.

2) Evaluate the suitability of ten different pollinator friendly marketable crops (cut flowers) for inclusion in pollinator habitat. To accomplish objective #2:

We will use a randomized block design with 10 treatments and 3 blocks.

Treatments will be ten economically viable plant species (cut flowers) with three repetitions. Flowers chosen for the secondary study include:

• Achillea millefolium 'Red Velvet'
• Achillea millefolium 'Terra Cotta'
• Campanula glomerate 'Superba'
• Helenium autumnale MARIACHI 'Bandera'
• Echinacea 'Hot Summer'
• Liatris spicata 'Floristan Violet'
• Liatris spicata 'Floristan White'
• Penstemon digitalis 'Husker's Red'
• Salvia nemorosa 'East Friesland'
• Veronica longifolia 'Allord'

• Blocks control for time of insect counts. We can record insect visitors in ten plots at once.
• Plots will be 3 x 2 ft. with 5 plants placed 12 inches apart.
• Paper and bark mulch will be applied in early spring to prevent winterkill through crown rot.
• Pollinator visitation and diversity will be assessed on at least six full days within each treatment spaced from April through August using time lapse video camera surveillance.
• Time lapse photography will be scanned and pollinator and beneficial insects will be counted and identified. To conserve pollinators, most pollinating insects will be identified in a noninterfering manner through photos. Periodic scouting of the treatments will allow net capture of unknown insects for positive identification.
• Plant vigor and growth will be accessed each month from April through August through foliage area cover and flower numbers. Due to the extreme vigor of the plants, this was done by recording percentage plot coverage and bloom coverage of each plant variety. 
• Treatments will be analyzed for significance using ANOVA.

3) Develop an easy to establish, non-invasive, cost-effective and obtainable pollinator and beneficial habitat modules for western growers. We will use information from our research, as well as other expert resources, to design pollinator habitat modules specifically for the western US that can easily be used singularly or combined as strips or patches by growers.

4) Survey of pollinator and beneficial diversity and abundance on a Washington certified organic farm. Information we obtain from insect counts will be used to compile a survey of insect diversity and emergence during the growing season. 

5) Encourage growers to put in more farmer friendly pollinator and beneficial insect habitat. This will include compiling information, taking plant and insect photos, and creating educational material specifically for growers on sustaining native pollinators and beneficial insects through low care, non-invasive, cost-effective, and economically viable pollination modules.

6) Write research paper on study. Analyze data and write up findings.

Project Update for 2018

Plant selection and care

Criteria

A lot of debate went into plant selection. Jane Fluegel, our horticulture expert, eventually selected a great group of plants. We decided to use only economically valuable cut flowers for the experiment. Our criteria were that all the plant species selected needed to:

1. be attractive to diverse native pollinator species
2. have a beautiful long-lasting bloom suitable for farmers to sell as cut flowers (some of our choices reflect the growing trend of more 'natural' appearing flower arrangements)
3. enjoy a long blooming time
4. not be a noxious weed or a problem for local farmers
5. have robust growth without an invasive tendency
6. be locally available; we didn’t want rare or hard to obtain flowers
7. thrive in the Pacific Northwest climate without being babied

Flowers species chosen for the study:

For the main study on the effect of plant architecture on pollinators we paired similar plants:

1. Salvia Caradonna (S) with Nepata x faassenii ‘Walker’s Low’ catmint (C)
2. Leucanthemum x superbum ‘Alaska’ (L) with Gaillardia ‘Burgundy’ (G)

As described in materials and methods we will test ten variables with five repetitions and four plant species: SS, CC, LL, GG, SC, LG, SG, LC, SL, and CG

Flowers chosen for the secondary study to valuate the suitability of ten different pollinator friendly marketable cut flowers for inclusion in pollinator habit:

1. Achellea millefolium 'Red Velvet'
2. Achillea millefolium 'Terra Cotta'
3. Campanula glomerate 'Superba'
4. Helenium autumnale MARIACHI 'Bandera'
5. Echinacea 'Hot Summer'
6. Liatris spicata 'Floristan Violet'
7. Liatris spicata 'Floristan White'
8. Penstemon digitalis 'Husker's Red'
9. Salvia nemorosa 'East Friesland'
10. Veronica longifolia 'Allord'

We obtained 3.5” potted plants in mid-May. We immediately repotted all the plants in gallon pots using organic potting soil. Plants were cared for and watered during the hottest part of summer in preparation for fall planting. This is where Jane's expertise in nursery plants came in handy. Since we started plants early, we will have large healthy two-year-old plants next year. See the size of our plants in the photo to the right (photo of Susan Fluegel with some of our healthy and beautiful potted plants in summer of 2018).

After selecting our plant species, I realized that I had underestimated the space needed for mature plants of these large robust species. Since the plot size needed to stay small enough to position a camera over the plot and care for plants without stepping into the plot and compacting soil, I consulted Jane and adjusted the number of plants per plot. We reduced the number of plants in each plot to six plants per plot for the main experiment. The reduction in plant number per plot allowed us to use additional plant funds for the supplies and labor needed to repot plants and grow flowers to a gallon size. With this change to much larger plants we will get earlier and larger blooms and more pollinator data the first year.    

Field and plot preparation

The field area we are using for the study had slightly compacted soil. It was covered with grass, a few Canadian thistle patches and some other hard to kill perennial weeds. From previous experience with defeating these troublesome green foes, we knew that solarization was a great way to control perennial weeds organically.

The plot area was covered with heavy tarps for 2 months in the summer to kill the perennial weeds through solarization. After weeds were killed, the soil was spiked to reduce compaction, plowed, and rototilled using a small 1950’s John Deere tractor. 

During tillage the soil was amended with added horse manure and bark mulch to raise organic matter content, add nutrients and improve soil water holding capacity. This step gives plants a great head start. Both the spiking of the soil and the plowing in of amendments took us a little more time than we estimated but the great results were worth it.

The plot area was measured, the plot was squared, and individual plots were marked using 2-foot wooden surveyor stakes. Due to the number of plants we selected per plot, and for ease of drip irrigation installation, the plot size ended up being three feet by two feet.

Experiment planting

Large, fully rooted gallon-size plants were planted one foot apart in the marked plots in late September 2018. Susan, Jane, and our youngest farmers, Ollie and Annabelle, helped plant. Plants were lightly mulched for winter with bark.

The light mulching was to avoid crown rot caused by excessive winter moisture around the base of new plantings. We wanted to be cautious since this can be a major problem with fall plantings in the Pacific Northwest. Our rainfall is greater in the winter months than the summer months.

All the plots will have organic weed paper placed around the plants and full bark mulch as discussed in the grant placed down in early spring before weeds start to emerge.

Plants had ample time to root before winter dormancy. As of early January 2019, experimental plants look healthy and viable. Any plants that winter kill will be replaced in spring (we purchased a few extra plants for insurance). The replacement plants were left in their pots, heeled into the soil in a sheltered area, and mulched heavily with leaves. 

Winter 2019 

Luckily a thick snow covered the plants during the extreme cold of early 2019! This photo illustrates the mood perfectly as we waited for summer hoping that our little plants were safe under a blanket of white.

Winterkill is always a possibility with fall planting but an insulating blanket of snow makes it much less likely. Usually it is not the cold that kills or harms plants in winter but the freezing and thawing cycles. Snow cover and/or mulch helps by insulating the ground and preventing wide temperature swings.    

Camera mounting: planning, prototypes and execution 

Initial planning winter 2019

Originally, we planned to mount the waterproof time-lapse cameras on permanent fence posts next to the plots. After some discussion we decided to build wooden camera mounts so that the cameras could be mounted above the plot. This would allow for a clearer view of the pollinators on the flowers and more accurate data.

Over the winter we built several prototypes of wooden mounts. They worked well but proved too heavy and ungainly to move easily. This was a problem since my mom and co-researcher, Jane, would have to move the cameras and I didn't want to make it too difficult for her or anyone else moving the cameras. I also wanted something that other farmer/researchers could replicate easily. Chris said he would help us by designing several different working camera prototypes to test. 

Working camera mount prototypes spring 2019: 

After field testing several camera mounting prototypes, including a lighter wooden frame and a simple step in fence post, Chris Bailey designed a lightweight sturdy camera mount for us. The winning design is shown at left.

These mounts can be made from plastic or metal pipes of any length. Our new camera mount is very light weight which allows us to easily move the mount. The pipe mount pole is stepped into the ground easily with a metal step and a bottom rebar spike. The rebar spike prevents the mount from moving or falling over in strong winds. 

We can also easily adjust the camera height from ground level to 6-feet tall. This was essential as the plants grew amazingly well and the camera had to be adjusted on a weekly basis during the spring and summer.

All of us are very happy with the custom made mounts. These camera mounts worked great and stayed upright even in wind or bad weather. Nobody's back got thrown out having to move them either!

In addition, Chris designed a small wooden shade to help protect the camera (which you can see in the upper right). The camera has a water proof case but it was starting to get extremely hot in the sun. The simple shade is held on by clips and helps keep the camera at a reasonable temperature. It does not influence the photos. The cameras can be easily detached from the mount with one step. Afterwards they can be taken out of their waterproof casing to change the batteries and memory cards. 

Update 2020:

These camera mounts still work great. When watching the pollinator video it was very obvious when we switched to these mounts. We eliminated the sway in the wind that we had with the first several mount designs. this allowed for easier viewing. 

In particular, incorporating a wood cover over the cameras helped protect the cameras from heat and rain. The ten cameras lasted in the harsh weather conditions for over two seasons. 

Late winter/early spring 2019

Almost all our plants survived the winter and look happy and healthy. We replaced the few plants that winter killed in the field with potted plants from the same age and batch.

Since the previous 2018/2019 winter was wet we were happy that we delayed mulching until spring. Organic weed paper was placed in all the paths and paths were mulched with several inches of bark. Low flow drip irrigation was installed between plants (3 drips per plot, 1 drip for 2 plants). Plots were mulched lightly. 

We have a weather station in the field to record temperature variation, wind speed, solar radiation, humidity and rainfall. Thanks to our weather station we now know that our field has some extreme temperature swings from day to night; sometimes over a 45 °F difference in one day! 

In spite of the wide temperature changes (dipping into freezing weather at night and jumping to a balmy 50-60 F during the day), the plants are looking great. The catmint and veronica are out growing the hot weather plants, gaillardia and shasta, at this stage.    

Due to some winter kill of the helenium autumnale MARIACHI 'Bandera' we added two catmint 'Walker's Low' to each helenium plot.  This made it a mixed plot of helenium and catmint. 

Late spring/summer 2019

Our decision to grow and use larger plants has been an outrageous success. The plants have taken off and are growing like weeds. In addition, planting in fall allowed for successful plant establishment and strong rooting which supported robust spring and summer growth. In areas with milder winters, fall planting will really jump start your perennials. We plan to recommend it as a possible strategy to establishing permanent pollinator plots. 

Some of the plants, such as the catmint and veronica, started blooming very early. We had been getting a lot of passerby interest in our pollinator plot but it really started to increase once the plants begin blooming. People were randomly stopping to take photos and ask about the project. We gave quite a few impromptu tours of the pollinator plot and discussed our research.

Insect diversity was very high. We attracted both native and non-native insects of all types including bees, flies, moths, butterflies and others. At first it was mainly native bees and other native insects. Later, in midsummer, three different hives of honeybees discovered the plot. We knew they were from different hives due to their different coloration. The honeybees gravitated to the catmint and often totally covered the blossoms. 

About midsummer, the noise from the pollinator plot in the middle of the day manifested as a very low key buzzing. It was very peaceful and relaxing. You could stand there and be surrounded by gorgeous flowers and hundreds of busy insects of all sorts. All the bees and wasps were very mellow and nonaggressive. 

Plant species influenced insect visitation; although all the flowers were visited throughout the day. You could confidently predict which flowers would attract the majority of some pollinators, however. For example, the catmint attracted more honeybees and large native bees while the shasta's attracted more small native bees and fly species.

Project Update for late winter/summer 2020

This was a very late spring following a cold winter without adequate snow cover. The catmint and salvia thrived and look wonderful. Both plant varieties are huge and the catmint grew like crazy.

Unfortunately, the gaillardia took a severe hit. In some plots the gaillardia was winter killed and in others it was stunted due to the cold spring. Since the gaillardia was missing or stunted, the other companion pollinator plants muscled into their space. In particular, the catmint took great delight in taking over any and all open space. By the time it got hot enough for the existing gaillardia to grow it was too late for it to compete. 

The shasta daisy had the opposite problem, some of the larger plants died back but all the plants reseeded themselves mightily resulting in a virtual sea of spring shasta seedlings. While this would be okay in a pollinator band or wild area, it is less acceptable in a research plot. Even with harsh and swift population control via hand weeding, the shasta overwhelmed the area. Since it reseeded the resulting plants were also less uniform and smaller than the original plants. They also bloomed later in the season due to smaller size. 

The plots still looked great and attracted attention (socially distanced attention mostly). Some of the plant varieties in our second experiment were spectacular. In particular, the liatris and Red Husker penstemon were gorgeous pollinator magnets. Like the gaillardia, the lavender did not withstand the cold winter. We replaced it with Campanula (bellflowers). 

We collected video data all season on all the plots, but due to these two problems could not analyze the second year of the first experiment. We were able to use and analyze the data from the plants in the secondary experiment though and that will be reported.   

In early fall, heavy smoke disrupted the pollinators for around a week. We recorded on those days but will not use that data.  After that pollinator numbers did not rebound.

Data Collection and Statistical Analysis

Plant time lapse videos were downloaded each day and stored in external hard drives. 

We watched videos for each plot to determine pollinator and natural enemy abundance and diversity. For consistency, we watched 30 minutes of video, from noon to 12:30 pm for each plot. Insects were counted using an area 1/4 of the total plot size. In plots with two species of plants, each of the plants were represented in the area chosen for counting. 

Video play speed was slowed down and insect numbers were recorded. In addition, each insect was identified according to family (or in some cases superfamily). This identification was made easier since we had spent time during the summer collecting, identifying, and releasing insects. It allowed us to visually identify our regular pollinators on camera later. 

Data was recorded.

Statistical analysis

Statistical analysis was done using JASP.  Data was analyzed with Repeated Measures ANOVA. Greenhouse-Geisser was used for sphericity correction. Time was the block (repeated measure) in the experiment. Plants were the variable. Data was further analyzed with Post Hoc tests using Tukey. Statistical details will be reported in the published paper.  

Charts and Graphs 

Charts and graphs were created on Microsoft Excel.