Thursday, August 7, 2008

PS 59-80: Monitoring the use of prescribed fire for oak savanna restoration at Sherburne National Wildlife Refuge in central Minnesota

Gary H. Swanson, U.S. Forest Service and Mark R. Fulton, Bemidji State University.

Background/Question/Methods  

The pre Euro-American settlement upland vegetation of Sherburne National Wildlife Refuge was dominated by scrubby oak woodland-brushland, oak barrens, savannas and prairies.  Beginning in the 1970s, prescribed burns have been used as a management tool for the restoration of savanna vegetation.  A program to monitor the success of prescribed burning in restoring oak savanna at the Refuge began in 1993.  Adaptive management based on results from such monitoring will increase the likelihood of success in this restoration effort.  This study focuses on three upland plant communities: dry oak forest (forest), oak savanna/oak woodland-brushland (savanna), and tallgrass prairie openings (grassland).  The main question is whether burn prescriptions and return intervals since 1993 have caused species composition in the grassland or forest communities to become more similar to the savanna remnants.  Non-metric multidimensional scaling and Mantel tests were used to examine trends in species composition and explore specific correlations between burning treatments and plant community changes.  Univariate analyses were also used to investigate trends in individual species and guilds.

Results/Conclusions

Savanna remnants existing on the Refuge today are shown to be a good, and possibly the best, model for restoration efforts available.  Mantel tests revealed significant correlations between burning and plant abundance, especially in the forest community.  Non-metric multidimensional scaling showed that, after a decade of prescribed burning, neither the forest nor grassland communities had become more similar in species composition to the savanna remnants.
Univariate analysis also revealed specific areas of concern.  Carex pennsylvanica increased in the forest community, and Poa pratensis increased in the grassland community.  These species may be stimulated by fire application at the wrong time of the growing season, and in the case of C. pennsylvanica by increased sunlight on the forest floor following each burn.  Tree regeneration in the forest and savanna communities is also a concern.  Pole trees have decreased by approximately 46% in both communities over the last decade.
It is concluded that other management treatments along with fire application will be necessary to accomplish savanna restoration goals at the Refuge.