1 FEBRUARY 2001
Contact: Alan Knapp
[log in to unmask]
785-532-7094
Kansas State University

Tallgrass prairies may provide early warning of climate change

MANHATTAN - Tallgrass prairies. One of America's most endangered
ecosystems.

While their size has diminished over the years - only an estimated 5
percent of the original tallgrass prairie in the United States exists
today - their importance in the ability to predict climate changes has
not, according to research conducted by two Kansas State University
scientists.

Long-term studies in tallgrass prairies may be able to predict and help
researchers better understand how ecosystems across North America might
respond to certain aspects of climate change, such as global warming;
droughts and changes in precipitation amounts, according to K-State
biology professor Alan Knapp. His research, co-authored by Melinda
Smith, a doctoral student in biology, is published in the Jan. 19 issue
of Science

Knapp and Smith compiled data from 11 long-term ecological research
sites across North America, selecting only those sites that had the best
and longest-term data on plant growth - known as production - and
precipitation. One of their goals was to determine which biomes, among
deserts, grasslands and forests, had more variable production; which
biomes responded most to precipitation variability.

"In this study, we were particularly interested in how different
ecosystems respond to variations in rainfall," Knapp said. "If rainfall
patterns change and variability changes in the future - as predicted by
climate models - how will that affect plant growth and which ecosystems
will be the most responsive?"

Knapp and Smith made a simple prediction. If the greatest precipitation
variability from year-to-year occurs in deserts and deserts are the most
water-limited biomes, then plant growth in deserts would be expected to
change dramatically from year-to-year; grasslands would be less variable
from year-to-year and forests would be the least variable of all.

Instead the data indicated their hypothesis wasn't supported at all.

"It turns out that what we found instead was that grasslands - in the
regions we live - have the greater year-to-year variation in plant
growth than in any other biome, even though rainfall variability isn't
as high here as it is in deserts, Knapp said.

That piqued Knapp and Smith's interest and led them to search for other
explanations.

"One of the important differences between grasslands and deserts is that
grasslands have a higher density of plants, and these plants can grow
faster than desert plants " Knapp said "If you think about a desert
environment, there's lots of bare ground between plants.

Knapp said even though regions with tallgrass prairies have less
rainfall variability, the ecosystem and its plants are well adapted to
respond to that variability.

"They grow very rapidly in wet years, yet can survive drought years as
well," Knapp said. "So we see the greatest variation in year-to-year
production in grasslands as opposed to deserts, and we see very little
year-to-year variation in production in forest sites. "

According to both Knapp and Smith, they were intrigued that their
analysis didn't fit their initial prediction.

"We knew the plant growth and precipitation are strongly related as you
move across North America," Knapp said. "We assumed that variability in
precipitation and variability in production would also be strongly
related across North America. It turns out they're not. One has to
factor in the characteristics of the biome as well, the potential
responses of the plants, how dense they are; how fast they can grow. You
have to factor in these biological characteristics with rainfall
patterns to explain these patterns of variation in production across
North America."

Knapp said although Kansans already have a strong affinity for the
state's tallgrass prairies, this research gives the prairies "added
value" in terms of their potential role in detecting changes in future
climates.


###
 http://www.eurekalert.org/releases/ksu-tpm013001.html