Harvest disturbances on forestland from 1600 to present

Linda A. Joyce1, David P. Coulson2, A.D. McGuire3,4, Richard Birdsey1, Brad Smith1, Todd Burnside4, Rose Meier4, 1USDA Forest Service, 2MATCOM, 3US Geological Survey, 4University of Alaska-Fairbanks

Influence of Historical Timber Harvest on Carbon Sequestration

Carbon sequestration in forests is strongly influenced by land use and management; yet until recently, ecological models examining the impact of climate change have focused on potential vegetation only. Our objective is to evaluate the impact of forest disturbances on carbon sequestration in US forests over time. The first step is to develop a nationally consistent data set describing the area of forestland harvested annually and remaining in forestland from 1600 to 1997.

Estimating Harvested Area on Forestland from 1600 to 1997

1952 to 1997: The limited available data on harvested area was combined with other timber inventory data to develop three approaches (Analytical, Adjusted, Statistical) to estimate the proportion of forestland harvested annually. The most inclusive nationally consistent data was for the 1980-1990 decade at the state level published by the USDA Forest Service. We compare these inventory estimates of the proportion of timberland harvested with estimates from our three methods in the graph below.

1600 to 1952: Little data on harvested area was available. We estimated the proportion of forestland harvested using two approaches: 1) linear interpolation from a 1952 estimate (models above) to first permanent settlement date, and 2) modify the interpolation in (1) with population estimates from date of first settlement to 1952.

Temporal Results for 4 States: Harvest Proportion from 1600 to 1997

Spatial Results: Annual Forest Disturbance from 1600 to 1997

Forest disturbance, proportion of forestland harvested annually, was estimated at the state level and then interpolated to the 0.5o latitude by 0.5o longitude grid scale. Shown here are the population-adjusted estimates of forest disturbance over time.

Modeling Land Use Shifts from 1600 to 1992

Land use changes were modeled as follows: annually, harvested area was computed by multiplying the harvest proportion by the available timberland (total forestland minus reserve forestland), area to be harvested was taken from primary and older cohorts first. Land was converted between forest and agriculture depending on the historical data for agriculture (Ramankutty and Foley 1999).

Forestland area declines with conversion to agricultural land use.

Model-estimated Forestland follows the Temporal Dynamics of Inventory-estimated Forestland Area (data from Smith)

Cumulative Area Harvested

Cumulative area harvested under the linear interpolation suggests that an equivalent area of the entire US forestland has nearly been cutover twice (1.94) whereas the population adjusted estimates suggest only recently has an equivalent US area been cut (0.88).

The two different harvest estimation procedures result in similar age class distributions through time but differ from inventory estimates in the young and very old age class.

Using the Two Different Disturbance Histories to Assess Cumulative Changes in Ecosystem Carbon Pools through Time

The linear interpolation of forest disturbance results in more carbon accumulation earlier than the population adjusted extrapolation, with little difference later.

Conclusions

We have established a method to estimate harvested area over time and link with land use change model. Modeled estimates of total forestland area follow the temporal dynamics of the inventory forestland estimates and are within 6-10%. The model produces a larger proportion of young and very old age classes than suggested by inventory data, although this comparison is confounded by uneven age stands in inventory data and the lack of fire as a disturbance in the model.

For more information, contact: Linda Joyce, ljoyce@fs.fed.us