Why study the impacts of invasive sweetclover on pollination?

            In Alaska, invasive plants have only recently moved off the human footprint into natural habitat. Warmer winters, longer growing seasons, and greater human activity may contribute to current rapid expansion across the state. Recently burned forest forms a major component of the vegetation of Interior Alaska, and this habitat is particularly vulnerable to invasion by early-successional non-native species. Fires are increasing in frequency and extent in Alaska. Melilotus albus (Fabaceae, sweetclover) was introduced to Alaska in 1913 as potential forage and has expanded rapidly along roadsides and more recently along floodplains and into burns.Sweetclover is highly attractive to pollinators and could disrupt the pollinator ecology of habitats in Alaska.

Melilotus could influence the pollination of subsistence berry plants...
            Two Vaccinium species (Ericaceae), Vaccinium uliginosum (bog blueberry) and Vaccinium vitis-idaea (bog cranberry or   lingonberry) form a major component of the subsistence lifestyle practiced across Alaska, both directly and by providing forage for other subsistence species  such as moose, caribou, snowshoe hare, ptarmigan, and grouse. Blueberries and cranberries overlap in habitat preferences with Melilotus. These species also share pollinators, including bumblebees (Bombus spp.), native solitary bees (Adrenidae, Halictidae, Megachilidae) and syrphid flies (Syrphidae).

Research Objective:
     Investigate the effects of white sweetclover (Melilotus albus) invasion on the pollination of Vaccinium uliginosum and Vaccinium vitis-idaea, and the subsequent effects on berry production

 Key Questions:

        • Does the presence of sweetclover change the pollen quantity and quality delivered to Vaccinium (berry) species?

Does the presence of sweetclover change Vaccinium reproductive success (fruit and seed production)?

        Does the effect of sweetclover on Vaccinium pollination and fruit set depend on how far apart the species are, or on how big the sweetclover patch is?

Does the presence of sweetclover change pollinator interactions with native plants?


We combined several different approaches to answer our questions:

    1.  Surveys:

  • We compared paired sites along the Elliot, Steese, and Dalton Highways in Interior Alaska. One site of each pair had sweetclover growing along the roadside, while the other did not.
  • We looked at pollinator visitation rates, and what pollinators visited the plants to construct pollinator webs
  • We evaluated pollen loads, fruit production, and seed set for Vaccinium species
  • We did experiments to figure out how many fruits and seeds the Vaccinium species would produce in the absence of pollinators.
Surveys have as advantage that they are realistic and can be done over large scales, but they have as a major disadvantage that we did not determine which sites got sweetclover and which did not. That means that the locations with sweetclover might be different in other ways as well - for example, they might just be better sites for plant growth in general. To deal with this we added a second approach:

2. Sweetclover supplementation experiments

  • We selected sites with lots of Vaccinium plants at the Bonanza Creek Long Term Ecological Research Areas.
  • At some of these sites we put sweetclover plants that had been grown in the greenhouse in the center of the site; at other (control) sites we did not add anything.
  • We marked Vaccinium plants (cranberries an blueberries) at different distances from these center points (where sweetclover had been added for the experimental plots), and then looked at the pollen they received on their stigmas (how much and of what species).
  • We tracked fruit and seed production on other plants at different distances from the central points.
  • Of course, we removed seeds from sweetclover as soon as they formed, and all sweetclover plants were removed and destroyed at the end of the experiment!
                                                                                                                                                                                                          Sweetclover plants ready to be planted

These experiments are an effective way of determining what the impact of the presence of sweetclover is (without any other consistent differences between sites), but they can only be done at a few locations. We think that the impact of sweetclover might depend on how much overlap there is in flowering times between the sweetclover and the berry plants. To evaluate this, we added a third approach:

3. Historical Data: Flowering Overlap Across North America
  • We located plant specimens of all three species at several herbaria across the US and Canada, noted the date of the collection, and scored the plants for the flowering or fruiting stage (phenological stage). These samples went back as far as the 1880s!
  • We are now using these data to construct models that tell us how the timing of flowering and fruiting changes depending on the geography (for example, latitude and elevation), and how well we can explain this by the climate at those sites
  • This will allow us to predict where flowering is likely to overlap in Alaska (and elsewhere), and by how much. It also allows us to determine if flowering times have changed over the past 100 years, and whether they are likely to continue to change in areas where the climate is warming.
Historical data allow us to look back in time, but they are less reliable than data collected with known protocols. So we have added one further component:

4. Current Phenology Data: Citizen Scientist Project
  • We asked people across the state and across the continent to help collect data on when our three species of interest produce flowers and fruits.
  • This complemented the historical data, and helped us to test whether our models work: if they do, our predictions should match what people are observing in different locations!
  • Over 200 volunteers contributed data to this citizen science effort! For information click on the Citizen Science Tab!

Peer-Reviewed Publications from the Melibee Project

Spellman, K.V., C.P.H. Mulder, and  M.L. Carlson. 2016. Effects of invasive plant patch size and distance on the pollination and reproduction of native boreal plants. Botany 94: 1151–1160. [PDF]

Spellman, K.V., and C.P.H. Mulder. 2016. Validating herbarium-based phenology models using citizen science data. BioScience 66: 897–906. [PDF]

Spellman, K.V., A. Deutsch, C.P.H. Mulder, and L.D. Carsten-Conner. 2016. Metacognitive learning in the ecology classroom: a tool for preparing problem solvers in a time of rapid change? Ecosphere 7(8):e01411.10.1002/ecs2.1411.  [PDF]

Spellman, K.V., L.C. Schneller, C.P.H. Mulder, M.L. Carlson. 2015. Effects of non-native Melilotus albus on pollination and reproduction in two boreal shrubs. Oecologia 179:495-507.  DOI: 10.1007/s00442-015-3364-9

Spellman, K.V. 2015. Educating for resilience in the North: Building a toolbox for teachers.  Ecology and Society 20(1):46. [PDF]

Bestelmeyer, S.V., M.M. Elser, K.V. Spellman, E.B. Sparrow,  S.S. Haan-Amato, A. Keener. 2015. Collaboration, interdisciplinary thinking, and communication: new approaches to K-12 ecology education. Frontiers in Ecology and the Environment 13: 37–43. [PDF]

Preliminary Findings:

Overall Project
Blueberry and Cranberry Pollen and Berry Production
Cranberry Pollen and Berry Production over 2 years
Pollinator Webs
2012 Alaska Invasive Species Conference Presentation-Schneller and Spellman
2013 Alaska Invasive Species Conference Presentation-Mulder
2013 Alaska Invasive Species Conference Presentation-Spellman

Dr. Christa Mulder (UAF Project PI), Dr. Matthew Carlson (UAA Co-PI), Katie Spellman (PhD student), Laura Schneller (Master's Student), Dr. Jeff Conn (USDA ARS), Dr. Steve Seefeldt (USDA ARS)


Katie Spellman,
Mar 11, 2014, 6:51 PM
Katie Spellman,
Apr 11, 2012, 4:40 PM
Katie Spellman,
Nov 7, 2012, 12:27 PM
Katie Spellman,
Apr 11, 2012, 4:41 PM