This document…

Is a drafty initial step at understanding the ecological situation of the Driftless Region, Written in R-Markdown, and distibuted as an interactive HTML file we hope it promotes discussion and is through provoking. As a “rapid assessment” it is not a survey and does not include all relevant literature or datasets. That said, we feel that the general patterns illuminated here will help with priority setting and knowledge gap identification moving forward.



The Driftless Region

The setting in prose

Each morning, the sun rises over the Oak savannahs and prairies of the Driftless region. The sunshine envelopes moss-covered bluffs and rocky ridges, unscathed by land carving glaciers. Located in south-western Winsconsin, south-eastern Minnesota, north-eastern Iowa, and the very north-west corner of Illinois, the Driftless region is home for a variety of unique flora and fauna.

Waking in the warmth of the morning sun, various small mammals like squirrels, mice, and voles scuttle amongst fallen oak leaves, munching into seeds and acorns. Many birds use the prairies for nesting including eastern kingbird, indigo bunting, and eastern bluebird.Larger fauna like White-tail deer, elk, and coyote wade through the tall prairie grasses and wildflowers. Grasses include big and little blue-stem, Indian grass, northern drop-seed, purple love grass, and numerous other species. Some forbs include pasque flowers, lead-plants, purple prairie-clovers, yellow coneflower, black-eyed susan, sunflowers, goldenrods, and asters. Due to the presence of both grassland and woodland, Oak savannas contain far more biodiversity than either of the two alone.

The rugged beauty of the Driftless region attracts a variety of nature-lovers, hikers, bikers, and fishermen alike.

There are few natural areas in the world left untouched by man. Like much of the land on Earth, the area we call the “Driftless” region has seen drastic changes over the course of recent history. Located in south-western Wisconsin, south-eastern Minnesota, north-eastern Iowa, and the very north-west corner of Illinois, here we explore the transformation of nearly sixteen million acres of the Driftless.

“Here, without knowing it, we still were at school, every wild lesson a love lesson, not whipped but charmed into us. Oh, that glorious Wisconsin wilderness!” – John Muir.

The setting in numbers

The Driftless region is ~24,000 square miles (~15.4M acres) with little evidence of glaciation. The area was “shielded” from the latest glacial activity by multiple factors including (Shimek 1948, Braun 1950, Hartley 1957,1966): * the “Watersmeet Dome”, or an uplifted area of cambrian rock that was layered over basalt (northern WI and western Upper Peninsula of MI) * bedrock of the Door Peninsula (eastern WI) * large amounts of snow dropped on the area

The resulting landscape is an eroded plateau with deeply dissected river valleys.

Since the presence of recent European fire-supressing settlers, the ecosystems of this area have changed over time.[add some ecological facts about importance of fire??] In this project, we take a look at data supplied by The Nature Conservancy’s LANDFIRE team to analyze how the Driftless area may have changed over years of intentional fire suppression.

Historic Vegetation Patterns

We are not aiming for the past, but rather a future landscape that can weather climate change, invasive species and other (perhaps currently unknown) threats in a way that preserves ecosystem services such as pollonator habibat, water filtration and storage, carbon sequestration, recreation and forest products. While we are not looking to restore the past, looking at historic locations and functionining of ecosystems can help identify where conservation might be most successful, how to increase resiliency and relative importance of management regimes (e.g., prescribed fire).

Here we map and chart “grouped” LANDFIRE Biophysical Settings, which represent “the vegetation that may have been dominant on the landscape prior to Euro-American settlement and is based on both the current biophysical environment and an approximation of the historical disturbance regime” (from https://www.landfire.gov/bps.php). Grouping was done in this rapid assessment to help simplify review and data presentation.

Map of Historical Ecosystems (Grouped LANDFIRE Biophysical Settings)

The patterns mapped here largely match expecations based on earlier work by Prior (1991 and Albert (1995) which was based on the General Land Office (GLO) surveys, then further refined by Shea et al. (2014). In the later paper, the authors also map out pre-settlement tree species (see Figures 3 and 4 in Shea et al. (2014)) reporting a dominance of bur, white and black oaks. They found that bur oak was strongly associated with low and medium topographic roughness; white oak with medium-high and high topographic roughness. Both were strongly associated with silt-dominated soils. Black oak was associated with sandy soils and medium-high roughness.

LANDFIRE Biophysical Settings data was developed by modeling relationships between ecosystems and biophysical paramters (e.g., climate, surficial geology, topography) and incfluenced by a review of GLO survey data, so similarities were to be expected.



The dominance of fire-resistant oaks, and extreamly fire-adapted prairie suggests a landscape that experienced a substantial amount of fire. On average the Oak and Pine ecosystems had a mean fire return interval (MFRI) ranging from 5-24 years and the prairie systems had a MFRI of ~3 years (including all embedded Biophysical Settings and their succession classes). Based on our maps of historical ecosystems, and modeled disturbance regimes from LANDFIRE, we calculate that the region had ~2.8 million acres of fire per year!





Current Vegetation Patterns

The Driftless Region is still beautiful, but very different from how it was just a few hundred years ago (again, see Shea et al. 2014, plus Knoot et al. 2015). Changes include, but are not limited to:





Currently the land is dominated with agricultural land use and mesophytic ecosystems illustrated here in a map and a chart.



Past vs. Present

Of the roughly 16M acres that are the Driftless Region 11+M acres (~70%) have been converted to agricultural or developed land use. Of the remaining unconverted lands some have shifted in type. For example, over 865k acres if what was mapped as Oak and Pine systems is mapped as Maple-Basswood today, presumably due to fire suppression. To explore patterns of change interact with the chord diagram below.

Chord diagram of ecosystem changes

To visualize patterns of change between the historical and current ecosystems we combined the LANDFIRE Biophysical Settings and Existing Vegetation Type datasets, then “cleaned”, or grouped the vegetation types into similar categories. We then used that data to generate the interactive chord chart below.

Hover over the outside ring segments. For example, if you hover over the darkest green segment of the “Agricultural or Developed Land Use” you will see which ecosystems, and how much of each has been converted. We find that most of what is now Agricultural or Developed Land use came from areas formerly mapped as Oak and Pine. Conversely, if you hover over the outside grey segment of Historically Pine and Oak Systems you’ll see that very little of that type remains, and that a fair amount has shifted to Maple-Basswood.




What is the current vegetation cover like in the main ecosystems of the Driftless Region?

Vegetation cover is a defining characteristic of an ecosystem. That is, for any ecosystem how much of it is herbaceous or tree vegetation, and how dense is the vegetation?

We can compare current vegetation cover to estimated historic, or reference cover to get a sense of how structurally departed an ecosystem is from natural conditions, and what we might need to do about it. For example, for many eastern oak ecosystems if the canopy cover is too great, then the native diversity of understory vegetation declines, as does regeneration of oaks. If canopy closure is too high, and if “desireable” tree species are in the understory, then thinning might be advised.

Methods

We can explore potential ecosystem distribution and current cover conditions by combining LANDFIRE Existing Vegetation Cover and Biophysical Settings (historical ecosystem) data for the region of interest, in this case the Driftless Region of the Midwestern United States. We did this combine in ArcGIS, the exported the resulting attribute table to Excel.

To make exploration easier (and arguably so that it makes more sense) we first removed agricultural and developed lands, then combined multiple LANDFIRE BpSs into “Groups” in an Excel pivot table. For example, we combined 7 BpSs into the “Oak and Pine Systems” group.

We then created the “sunburst” plot below in R using the "sunburstR’ package.

Explore

In the sunburst chart below you can explore the canopy cover per ecosystem patterns by first hovering over the inner ring to get the percentage of each ecosystem, then moving your cursor outward to explore canopy closure per ecosystem. Percentages will appear in the middle of the chart, and the “path” (e.g., “Oak and Pine systems > Tree Cover >= 70% and < 80% = 17.6%”) will appear to the top right.

The canopy closure percentage colors stay consistent from ecosystem to ecosystem, and add up to 100%.

Legend

Observations and limitations

Of the remaining natural lands the rough distribution of ecosystems is as expected based on LANDFIRE Biophysical Settings (i.e., historical ecosystems) data, and canopy closure is dominated by tree cover greater than 80% and herbaceous vegetation (see note below).

More specifically:

  • The Oak and Pine Systems are most dominant, covering roughly 50% of the Driftless Region
  • Tree cover is greater than 80%, presumably (and problematically) due to fire suppression in the fire adapted ecosystems (e.g. praries). This would be normal in the riparian ecosystems.
  • The amount of herbaceous vegetation in the non-prairie ecosystems raised red-flags. Upon further inspection we learned that these areas are classified as “ruderal” vegetation. As such they may or may not be areas worth further inspection for potential restoration.
  • The base ecosystem map represents historical (or potential). While we removed any current agricultural or developed areas, we did not assess ecosystem conversion here, which could be a conservation issue (e.g., oak types converting to more mesophytic types). As such, this is only one part of a full ecosystem assessment.

Limitations

The LANDFIRE Biophysical Settings (representing historical) and Existing Vegetation Type datasets were generated using very different methods and are designed for use over large landscapes. These visualizations are for exploring patterns and for identifying potential future research.

Resources

Albert DA (1995) Regional landscape ecosystems of Michigan, Minnesota, and Wisconsin: a working map and classification. General Technical Report NC-178. USDA Forest Service, St. Paul, MN, 250 pp

Prior JC (1991) Landforms of Iowa. University of Iowa Press, Iowa City, IA 168 pp