“Right plant, right place” has long been a mantra in the landscape design world. But what about “right plant, right pollinator”?

Pollinator Companion, a new interactive plant-selection tool available through CNPS’s Calscape website, identifies native companion plants to support native pollinators, including those important to crop production. For example, squash bees (genus Peponapis) excel at pollinating plants in the Cucurbitaceae family, such as zucchini and pumpkins. By creating specific habitat for the bees—say, by planting vinegarweed (Trichostema lanceolatum) or Missouri gourd (Cucurbita foetidissima)—farmers and gardeners can boost pollination support for their current crops while addressing the issue of pollinator decline.

Ultimately, the goal of Pollinator Companion is to reverse pollinator decline by encouraging people to plant high-impact pollinator habitats. Even small-scale patches of habitat can make a difference by providing corridors for native pollinators to move through the landscape. CNPS, together with researchers collaborating through the National Center for Ecological Analysis and Synthesis’ (NCEAS) Morpho Initiative, aims to facilitate better-informed planting decisions through tools like Pollinator Companion.

“One limitation of a lot of plant selection resources is that they don’t really have the type of spatially resolved information that we need to confidently say, ‘This plant is going to support this pollinator in this specific ecoregion or habitat,’ ” says Dr. Chris Cosma, a researcher at the Conservation Biology Institute in Corvallis who co-leads the Morpho Initiative working group.

You have thousands of species in California, both for plants and for pollinators. I think this combines the best of my worlds, which is using hard-to-deal-with data to hopefully solve real world problems.” – Dr. Laura Melissa Guzman

The idea for Pollinator Companion was born from Cosma’s doctoral research at University of California, Riverside. There, he built a tool called Butterfly Net, using ecological network methods to create ranked geospatially informed planting recommendations for native companion plants that host butterflies and moths. Benefits of this type of companion planting may include deterring pests, improving soil conditions, and
attracting beneficial insects.

While Cosma was working on Butterfly Net, Dr. Laura Melissa Guzman was independently working on a similar tool at Simon Fraser University in British Columbia. In a moment of serendipity, the two researchers met when Guzman gave a seminar at UC Riverside, and they decided to collaborate. Both researchers are now based out of state; Cosma works and lives in Washington, and Guzman is an assistant professor at Cornell University in Ithaca, N.Y. However, contributions from their fellow collaborators at CNPS—including Andrea Williams, former director of Biodiversity Initiatives, and Calscape product manager Jessica Woodard—as well as funding from the California Department of Food and Agriculture (CDFA) prompted them to develop the tool in California. The state’s abundance of data on native plants and pollinators (not to mention the potential to reach millions of users by integrating the tool within Calscape) only reinforced that decision. Led by Cosma, Guzman, and Woodard, scientists with the Morpho Initiative built a data-driven model that forms the backbone of the Pollinator Companion.

“I love data,” Guzman says. “I love heterogeneous and hard-to-deal-with data sets. And this actually presented a really interesting data problem because you have thousands of species in California, both for plants and for pollinators. And so, for me, I think it combines the best of my worlds, which is using hard-to-deal-with data to hopefully solve real world problems.”

The model distills data from a whopping 802,303 pollination and herbivory observations spanning centuries. The data was sourced from the Global Biodiversity Information Facility (GBIF), which collects records as varied as museum specimens dating back to the 1800s and smartphone photos from applications such as iNaturalist. On its own, the GBIF data was unmined gold, not readily available to or easily parsed by the public. Pollinator Companion enables users to design habitats based on this treasure trove of data. The model predicts whether interactions between native plants and pollinators are likely to occur in a specific location and at a specific time of year. Additionally, the data that feeds into the model includes the geographic distribution of native plants and associated pollinators, the bloom time of the native plants, and the active period of the pollinators.

Accounting for gaps in the available data, the model can extrapolate whether plant-pollinator interactions are likely to occur in a location or timeframe lacking data, based on nearby data points. For example, if the Western bumble bee was observed on California buckwheat in Fresno in May and on the same plant in Sacramento in July, the model would report a high probability of observing the Western bumble bee on California buckwheat in Modesto in June—even with no prior observations. Backing from CNPS and CDFA has enabled the researchers to develop their statistical model into the interactive Pollinator Companion tool.

While bees are by far the most efficient pollinators of most crops, the researchers sought to represent other important native pollinators in the tool as well. “Most of the time, we have incomplete information about the full suite of pollinators that are actually important for our crops,” Cosma says. Even though the most effective pollinators may belong to a specific group, total crop yield increases with a more complete or diverse pollinator community. “Moths in particular are a great example of a group of pollinators that we just don’t have as much data for,” he adds. “But that’s mostly because we’re not awake at night studying them.”

Plans for the Pollinator Companion include adding other underappreciated pollinators like beetles, wasps, and more types of flies. Calscape will integrate the pollinator data throughout the website, where it will be available on plant profiles and even a new pollinator search. The research team also plans to conduct two validation studies: One study will test their model in a field experiment to see if their predicted plant-pollinator interactions occur in the real world. The other study will test whether people make better planting decisions when using their interactive tool, compared to a static planting guide. Eventually, the researchers say, they hope to expand Pollinator Companion nationwide.

“We really want this to be a tool that facilitates these sorts of pollinator habitat design decisions by the general public and professionals as well,” Cosma says. “We want to connect people with the evidence and the ecological data in a user-friendly way.”

Pollinator Companion is currently under beta testing. You can contribute to the development of Pollinator Companion by using the tool and providing feedback. Visit calscape.org/support-wildlife for more information.