Fire and dynamics of granivory on a California grassland forb

Summary

There is evidence that competition by invasive exotic annual grasses reduces fecundity of large-flowered fiddleneck Amsinckia grandiflora, an endangered herb native to California grasslands. In California, fire is sometimes used to control exotic annual grasses and to remove plant litter to stimulate native plant growth but whilst fire may have beneficial impacts by reducing competition from exotics, there may be other consequences for native plants. This study examined the effects of prescribed burning on subsequent granivory of A.grandiflora seeds by e.g. rodents and birds, and post-burn reproductive success.

Study area: The study was undertaken in grasslands over five years (1998–2002) in the Altamont Hills, California, south-west USA. Experiments were undertaken within two areas of a reintroduced population of A.grandiflora sown in 1993 growing less than 500 m from a native population.

Area 1 - a 25 x 30 m area of five blocks with 10–14, 80 x 80 cm plots in each.

Area 2 - adjacent to Area 1, comprising five blocks with four 2 x 2 m plots within each.

Plot design: Twenty-five 9 cm long galvanized nails were pressed into the soil flush to the ground in a 5 x 5 cm grid in the centre of each plot. A single nutlet (seed) was fixed to each nail head with double-sided sticky tape. A A.grandiflora plant typically produces fewer than 25 nutlets/year. Thus the density of seeds used was considered to resemble densities that occur naturally.

Treatments and experiments: In 1998 and 1999, both burned and unburned plots had netted (to exclude birds) and no-exclusion (open) treatments.

1998
On 11 June, half of Area 1 was burned and 4 days later five nutlet grids (1/block) were set in each of four grassland treatments: netted+burned, netted+unburned, open+burned, open+unburned.

Two rounds of experiments were conducted in 1999, 2000 and 2001.

1999
Round 1 - set out on 26 April at the time of A.grandiflora senescence. Five replicates in Area 1 (one in each block) were placed in each treatment=cover type: netted+disturbed (burned previous year), netted+unburned, open+disturbed (burned previous year), open+unburned. Observations of seed granivory were made over 5 weeks.

Round 2 - Burns conducted on 28 June and seed grids set out that day. Observations of seed granivory were made over 3 weeks.

2000
The aim of these trials was to determine the effect of rodent trapping on granivory. Five replicate plots were established in open+unburned locations in Areas 1 and 2 (one replicate per block) which were to be trapped, and in an untrapped control area 10 m away.

Round 1 –. No burns took place in 2000 and no netting was installed. Commencing 1 May 2000, observations of seed granivory rates were made over 3 weeks. At the end of round 1, lethal snap traps were installed in Areas 1 and 2. The traps were installed in a grid where traps were spaced 5m apart. Traps were baited and checked each morning.

Round 2 - On 5 June, the traps were removed and the experiment was run in different plots in the same treatment areas. Observations of seed granivory were made over 3 weeks.

2001
Round 1 - On 27 April 2001, 10 plots (5 in Area 1; 5 in Area 2) were established in open+unburned locations. Observations of seed granivory were made over 3 weeks.

Round 2 - On 20 July 2001, after a selective burn of three of the four plots in each block in Area 2, 10 plots were established in Area 2 only. Five plots were located in unburned plots and five were in burned plots. These were checked over 3 weeks. On 3 August, all burned plots were restocked with nutlets, an additional plot was added, and three of the six burned plots were netted; unburned plots had little seed loss over these 2 weeks and were left as they were. Observations of seed granivory were made over 2 weeks.

2002
Round 1 - On 8 May pre-burn plots were set out comprising five plots in Area 1 and 10 in Area 2 (five disturbed and five unburned). Observations of seed granivory rates were made over 3 weeks. One plot per block was burned in Area 2 on 20 June.

Round 2 - On 1 July post-burn plots were set out, comprising five open+unburned plots in each of Areas 1 and 2 and five open+burned plots in Area 2. Observations of seed granivory were made over 4 weeks.

Seed losses were high (51–99%) and final granivory rates were generally similar among treatments. Nearly all seed grids were discovered (>10% seed loss) by granivores in all trials in all years. There were significant differences in seed loss among years, within-year effects due to burning or bird exclusion were much less apparent. Fire affected granivory when granivory rates were low, but when levels were high (as in 1998 and 1999), it appeared not to affect granivory rates.

In 2000, seed loss in the unburned, open plots were lower than in previous years (14%), probably as rodent numbers were low; trapping yielded only a one animal. In 2001, burned/open plots experienced significantly more granivory (87%) than either burned/netted plots (37%) or unburned/open plots (47%), and in 2002, granivory (100% seed loss) was high in the burned, open plots.

It was suspected that burned areas would suffer more granivory than unburned areas because of greater visibility of the nutlets and fewer other seeds. It was also thought that rodents would be less likely to forage in open areas created by burning, thus birds would be the main granivores in the open burned plots. However, in 1998–99 granivory was no higher in open, burned plots than in netted, burned plots, and granivory was no higher in open, burned plots than in open, unburned plots. The lack of difference between netted and open plots indicated that birds were therefore not the primary granivores, at least in these years.

Ants were responsible for loss of some nutlets (ants were observed removing nutlets from nails), but the authors consider that due to the spatial scale of the experiment it was unlikely that ants were responsible for the majority of the granivory occurring across all plots.

Lack of difference between unburned and burned plots indicated that lack of cover neither, enhanced or reduced granivory. If rodents are the main granivore of A.grandiflora seeds, high granivory rates in burned plots in 1998–99 could be explained by high rodent numbers. It is also possible that the rodent species composition changed over the course of the experiment, with different species having different foraging responses to changes in plant cover, but this was not investigated.


Note: If using or referring to this published study, please read and quote the original paper, this can be viewed at:

http://www.springerlink.com/content/l5q1365ql70k541j/fulltext.pdf