Within the northern Great Lakes region, mesoscale (10s to 100s of km2) forest patterning is driven by disturbance dynamics. Using original Public Land Survey (PLS) records in northern Wisconsin, USA, we study spatial patterns of wind and fire disturbances during the pre-Euroamerican settlement period (ca. 1850). Our goals were: (1) to determine how effectively wind and fire disturbance can be econstructed from the PLS, (2) to assess the roles of wind and fire in shaping vegetation patterns, (3) to evaluate landscape to regional controls of wind and fire regimes, and (4) to assess the potential for interactions between these disturbances.

Our analyses indicate that only relatively intense fire and wind disturbance can be reliably detected from the PLS (62–68% canopy removal). Heavy windthrow was more prevalent than fire disturbance in presettlement forests, and wind-disturbed patches were comparatively smaller and more complex in shape. Disturbance rotation periods ranged

between 450 and 10 500 years for heavy windthrow and between 700 and 93 000 years for stand-replacing fire. Occurrences of wind and fire disturbance were related to geographic province and to regional soil patterns; analysis further suggests a negative interaction between the two disturbance types. Given that severe wind disturbance was infrequent, mature to old forests of late-successional species dominated much of pre-Euroamerican northern Wisconsin, but wind disturbances may have allowed regional persistence of less shade-tolerant species, such as Betula alleghaniensis. Pine-dominated vegetation was limited to regions with more frequent fire, but frequencies of stand-replacing fire derived from survey records were insufficient to maintain these successional vegetation types; we suggest that frequent surface fires, not recorded in the PLS, along with infrequent stand-replacing fire, maintained these vegetation types.

The extensive nature of the PLS provides a powerful baseline for addressing changes

in forest conditions and disturbance regimes associated with climate and land use for both

the present and more distant past. Such baselines are informative in discussions of historical

variability and restoration silviculture.