We are the grad students from the Melibee project team. Meli, for Melilotus albus/officinalis or sweetclover, and bee, for the potential shift in pollinator services to native plants that sweetclover may be causing. We are particularly interested in plants important for subsistence food in Alaska, our wild berries.
1
2
3
Pollinators are vital to flowering plants. In our case of sweetclover, blueberry, and cranberry, pollinator exclusion experiments in 2010 showed that these three plants are dependent upon insect pollinators to produce fruit. The insects move pollen between flowers and between plants, allowing plants to exchange genes for a new generation of plants. Both plants and insects take advantage of this process. While the plants get pollen delivered, the insects take nectar and pollen for food and nest provisioning. This leads the plants to compete for the services of the insects, where the most attractive flowers get the most insect visitors. Plants use many strategies to be attractive, including flower size, shape, color, smell, amount of pollen, and amount of nectar.
4
The goal of attracting pollen‐carrying insects is to make fruits and seeds. For plants, this leads to a new generation of plants. For humans, this leads to berries that we can eat.
5
Our study took place in the boreal forest outside Fairbanks, Alaska. Sweetclover is attractive to insects, so we hypothesized that if sweetclover became present in the forest, it would affect the pollination and fruit set of the native berries.
6
Here is the schematic of how we set up our sites. The red dot indicates Melilotus transplants. Each green dot represents a focal plot. Plots are located in 5 orbits at different distances (from 1 to 40 m) from an experimental sweetclover patch. To observe pollinators, we randomly selected one focal plot at each distance category and observed pollinator activity. Insect studies utilized 4 addition sites and 4 control sites in 2011. On the plant side of the project, we completed two years of experiments on the plant end of things. In 2011, we sampled with 3 focal flowering blueberry plants (in a 3 m2 area) or 5 flowering cranberry plants (in a 1 m2 area) at each of the green dots in 6 control sites and 11 sites where we added 40 Melilotus plants. In 2012, we repeated the experiment with just cranberry, and added a higher patch size treatment. There were 6 control sites, 6 sites with 40 sweetclover added and 6 sites with 120 sweetclover in 2012.
7
We observed plant‐pollinator interactions by recording insect landings on plants within a 2 m x 2 m square.
8
We tagged the berry plants and tracked individual unopened flowers throughout the season until fruit set.
9
In addition, we collected stigmas, the sticky ends of the pistils, of non‐focal Vaccinium flowers at each plot, mounted them on microscope slides, identified the pollen grains, and counted the number of Vaccinium and non‐Vaccinium pollen grains on the stigmas.
10
The flowering plants and insects observed in the pollinator sites are used to construct the framework for the plant‐pollinator network diagram.
11
Plant‐pollinator interactions are represented by lines connecting the plants and pollinators. Thicker lines indicate more landings by pollinators on a particular plant.
12
This is the plant‐pollinator network diagram at control sites without white sweetclover.
13
This is the plant‐pollinator network diagram at sites with added white sweetclover.
14
15
Pollinator observations of non‐manipulated sites on the Dalton Highway showed similar trends as the manipulated sites near Fairbanks.
16
17
Because it is really hard to catch pollinator activity and get really good visitation rates from only a few short pollinator observation periods, and it is often difficult to tell if the pollinator is actually pollinating, or just touching down, we were able to turn to our pollen load data to get a better picture of which berry plants were actually getting pollinated. This graph shows us the difference in the % flowers visited between the sites with 40 Melilotus plants added, and the control site means at 5 different distance categories from the Melilotus patch or site center. It is like zeroing out a scale. The zero line represents the control sites means, and any dots above the line show and increase in the percentage of flowers visited in sites with sweetclover, or facilitative interactions, and dots below the line show a decrease, or competitive interactions. What we saw was a slight increase in the percent cranberry flowers visited directly adjacent to the sweetclover patch, and a 15% decrease in the number of flowers visited in the 8‐10 meter range. This competitive effect disappears as you continue away from the sweetclover patch.
18
These are similar style graphs, but showing the total number of pollen grains (pollen quantity) and the proportion of pollen that is actually Vaccinium pollen, or pollen quality. In both these graphs we see the same U‐shaped pattern as in the pollen quality and quantity found on Vaccinium stigmas. These are for Vaccinium vitis‐idaea, but the pattern was the same for blueberry, as well.
19
When we looked at the actual fruit set, we saw plants growing right next to the sweetclover were more successful at producing berries than at any other distance, and sweetclover had negative impacts on berry production for mid‐distances. Values are
presented as differences between Melilotus addition and control orbit means. +/‐ SE
20
We repeated the experiment this summer. Again, we saw a 15‐20% increase in fruit set for the cranberry plants close to the sweetclover at the 40 plant addition sites, but this year, we had increased fruit set levels extend throughout the further distances. Our climate data suggest that this change is due to climatic variation between years, with greater overall pollinator activity due to the increase in June sunny days optimal for pollinators. The real surprise, however, was that when we added 120 sweetclover plants to the site, we saw fruit production that was not that different from the control. We think there is a shift going on, from Melilotus being a pollinator attractor to get more pollinators to the site, to a total pollinator distractor when the sweetclover patch increases in number. We have some solid evidence to support this idea. Values are presented as differences between Melilotus
addition and control orbit means. +/‐ SE
21
Sweetclover has a complicated relationship with berry production, with negative, neutral, and positive effects on pollination and fruit set occurring all within a single site, and shifting from year to year. The impact of sweetclover on important subsistence berry plants depends on their distance from the sweetclover patch, climatic variables and the sweetclover patch size.
22
So what does this mean for our wild berry harvests? Any way you slice it, our data show that sweetclover is changing plant‐pollinator relationships in the system we studied. Sweetclover is unlikely to grow in the middle of berry patches in anywhere but the occasional low severity burn area, in which case, it tends to not persist and tends to have low flowering rates. In reality, sweetclover is a problem along long, linear disturbance features, like roads, rivers and trails. In these cases, the berry plants are not close enough to receive the pollination facilitation benefits that we observed at the 1‐2 meter distances. As a result, there may be a net negative impact to easily accessible berry picking habitat, with the effects neutralizing as you walk further from the road. One intriguing result from this year’s berry data, is that patch size is clearly important to understanding the impacts sweetclover can have on berry pollination. We only had the capability to add 120 sweetclover to our sites, but in many situations, like on the Dalton Highway (pictured here) or the Nenana River, you can have several thousand plants in a few meters. It is likely that this number of plants is an even stronger distraction for pollinators. Finally, timing is everything. The effects of sweetclover on blueberry and cranberry pollination in different parts of Alaska depends on the extent of the overlap in flowering times of sweetclover and berries. I’ll be talking more on this topic tomorrow, but if you cant wait to talk about phenology you should check out the poster by Steve Decina, one of our crew members on phenological overlap and non‐native species during the poster session.
23