.|  Baltimore Ecosystem Study
Vegetation Research Projects
 

 
Effects of a Changing Land Ecosystem on Terrestrial-Estuarine Ecology: History and Paleoecology
  • Grace S. Brush, Johns Hopkins University
The objective of this project is to reconstruct the mosaic of land use from forest to urbanization of the Gwynns Falls watershed and the effects of land use on the Patapsco River and Chesapeake Bay using historical records and the paleoecological record preserved in sediments deposited over approximately 2000 years. To achieve this objective, locations for sediment cores have been identified in the Baisman Run area. The cores will be dated using carbon-14 and pollen analyses and chronologies established. Pollen and seeds preserved in the sediment will provide a record of forest cover and land use that is being compared with historical records available since European settlement. The record of trace metals and nutrients preserved in sediment cores collected in the Gwynns Falls watershed are being used to determine the effects of different land uses on water quality. Historical records of developing urbanization are also being used for interpretation of changes in sediment cores. Comparisons of the effect of significant climate changes which occurred prior to European settlement on forests and water quality will be compared with effects of land use on water quality since European settlement.
 
Gwynns Falls Vegetation
  • Grace S. Brush, Johns Hopkins University
The Gwynns Falls watershed has changed from forest to agriculture to urbanization and suburbia over the last 250 years. The watershed is characterized by a geologic, soils and hydrologic gradient, ranging from schist that weathers into a well-drained soil in the upper watershed to saprolitic soils weathered primarily from gneiss in the lower section. Flooding is much less frequent in the lower watershed, where streams are more deeply incised. Superimposed on this gradient is a history of human activity that has altered local riparian habitats. Some of the vegetation distributions can be explained by specific human alterations of the natural gradients.
 
Different transformations of the landscape are overlain over time, resulting in changes that can affect processes important for nutrient cycling, species diversity, etc., such as increased runoff related to the building of roads and other impervious surfaces. Regenerating vegetation would be expected to respond to these changes, some of which are outlined in Table 1
 
DISTURBANCE DURATION EFFECT ON LANDSCAPE EFFECT ON STREAMS AND RIPARIAN ZONE
Charcoal production (lower 3rd) 1730 to 1810 1731 – Baltimore Iron Works formed deforestation row cropping probably less buffering of streams by trees, damming for furnace bellows, increased sediment supply
Agriculture began as early as 1664 peaked in 1900 declined in 1930s soil erosion from plowed fields; reduced infiltration; increased runoff; increased nutrients from fertilizers (post-1950) Floodplain vertical accretion due to increased sediment supply; floodplain less frequently flooded, wetland loss with agricultural drainage
Mill Dams first mill built in 1728; 1770s -numerous mills built to accommodate the Carribean wheat trade sediment trapped behind dams; decreased base flow below dams, permanent inundation above dams, fish blocks
Removal of dams 1914 – the majority of dams removed or breached sediment remobilized after dam removal to an unknown extent increased local, temporary sediment supply
Chrome mining (upper) 1820-1880 Local mine shafts Increased sediment inputs to streams, disturbance of riparian areas for placer mining
Urbanization (upper) 1957 and ongoing initial increase in sedimentation from construction; increased impervious surfaces; increased water and sediment runoff stream channels incised as sediment yield decreased and peak flow increased; stream banks increase in elevation

Table 1, Land use history in the Gwynns Falls watershed and its effect on the landscape and on streams and the riparian zone.

 
The Upland Vegetation
  • Grace S. Brush, Johns Hopkins University
  • Dan Bain, University of Pittsburgh
Measurements of the upland vegetation consisted of 71 random 100 m2 plots located throughout the upland part of the watershed, using random latitude/longitude points generated in GIS. These plots were then located on a geologic map in order to study the relationship of geology to vegetation distributions. In each 1002 m plot, all trees >= 2 and 2.5 cm diameter at breast height (dbh) were identified to species and measured. All stems <= 2.5 and 2 cm and >= 1 m in height were identified and counted. Four 1 m2 plots were established around the plot center, where percent cover of herbaceous species and species of woody seedlings were estimated using a 1 m2 frame divided into a grid of 100 10 cm2 cells. Species were designated as wetland or upland using the "National List of Vascular Plant Species" that occur in Wetlands (U. S. Fish and Wildlife Service Branch of Habitat Assessment 1996). Species were also identified as exotic or native using Gray’s Manual of Botany, 8th edition (Fernald 1950) and Gleason and Cronquist (1991). Basal areas (cm), number of stems, frequencies and importance values of the majority of tree species are given below.
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 
 

 
Tree distributions in the uplands are closely related to geologic substrates. Tulip poplar and red maple are the dominant trees on schist, gneiss ands marble – substrates with the highest water-holding capacity; beech and green ash on the less acidic amphibolite but with a lower water-holding capacity than gneiss and schist; red oak and black gum on the dry mafic substrates, and American elm and black locust on the Coastal Plain sand/clay/fill substrate. Below are histograms of species associated with one or two substrates. The following table is a list of species that occur on only one geologic substrate
 
schist  gneiss  amphiolite  marble  ultramafic  Qtzte, clay, fill
Acer saccharum    Chionanthus virginicus  Gleditsia trianthos  Quercus marilandica   
Asimina triloba    Fagus grandifolia    Quercus stellata   
Betula lenta    Gymnocladus dioica       
Castanea dentata    Magnolia tripetala       
Morus rubra    Pyrus melanocarpa       
Prunus pensylvanica           
Pyrus baccata           

 
In order to analyze the afforestation patterns taking place in an area with a progression of land uses beginning with agriculture and culminating with urbanization, the successional status of the species in the watershed was examined.
 
Each species was assigned a mean erodibility index number and mean percent forest cover based on the erodibility index number and percent forest cover of the plots in which they occurred. Each plot was then assigned to one of five successional stages (early, early intermediate, intermediate, intermediate-late and late) based on the species in the plot. As agricultural activity declined in the Baltimore region beginning about 1900 the first fields abandoned would most likely be marginal lands. These areas would undergo afforestation earlier and would at present be expected to have the greatest percent forest cover, and the largest trees.
 
Occurrence of different successional groups in plots arranged by average erodibility and percent forest cover.
 
successional stage (no. species) mean (range) erodibility mean (range) percent forest cover
early (6) 1.05 (0.4-2.1) 0.3 (0.2-0.4)
early to intermediate (5) 1.35 (0.5-1.9) 0.4 (0.3-0.6
intermediate (6) 1.5 (0.8-2.3) 0.4 (0.2-0.5)
late (18) 2 (0.9-4.4) 0.5 (0.1-0.8)

 
Early successional species occur more frequently on average in plots with a low erodibility and low percent forest cover than late successional species, although the range for late successional species is quite high. Succession after disturbance on the uplands of the watershed follows a normal successional pattern that is unchanged by a history of agricultural and urban disturbance.
The Riparian Vegetation
  • Grace S. Brush, Johns Hopkins University
  • Wayne C. Zipperer, United States Forest Service
Riparian corridors in the Gwynns Falls watershed follow a gradient from rural/suburban in the upper watershed to densely urban where the Gwynns Falls enters the Patapsco River. Before early to middle 19th century, the watershed was mixed forest and rural. Piedmont streams, which characterize most of this watershed, overflow their banks every two to three years. Consequently, growth and seed dispersal of plants occupying these areas are adapted to frequent, low magnitude floods. With the onset of urbanization, streams were modified to prevent overbank flooding, which resulted in stream incision with steep banks adjacent to the streams.
 
Riparian data were collected in 110 100 m2 (10 m x 10 m) plots located along 56 randomly selected transects that crossed 1 st to 4 th order streams. Transects were located on Maryland Geological Survey topographic maps and digital ortho-quarter quadrats, and were laid out perpendicular to the stream channel. The width of the floodplain was defined as the distance from the stream channel to a terrace or the change from natural vegetation cover to managed land cover (lawn, parking lot, etc.). For each transect, one plot was sampled on each side of the channel and every 30 m for the extent of the floodplain. Mean riparian width varied from 79.7±11.5 m in the upper part of the watershed to 46.2±10.2 m in the middle zone and 31.8±4.2 m in the lower zone. Relative elevations of each of the 45 transects were measured using CST/Berger 20x construction grade level surveying equipment. Relative elevations were measured approximately every 0.5 m or less where the elevation change was significant such as from the stream edge to the bank edge or old cut-off channels. The bank edge was defined as a significant break in slope of the cross-sectional profile from the stream channel to the floodplain. Elevation measurements were taken across the floodplain at 5 m intervals where the elevation change was not noticeable. The limits of the floodplain were determined by a significant change in elevation and vegetation. The depth of the stream was taken as the zero point in the cross-section.
 

Example of a transect across a floodplain
 

 

Sampling Design
 

 
Of 147 plant species sampled in the Gwynns Falls riparian corridor, 55 were exotics. Of those 55 exotics, 10 were trees and 45 were herbs. Diversity was similar throughout the corridor for trees and herbs, but higher for herbs.
 
  • Total number of species (147)
    • Trees (59)
    • Herbs (88)
  • Total number of exotic species (55)
    • Trees (10)
    • Herbs (45)
  • Total number of upland species (56)
    • Trees (23)
      • Native (17)
      • Exotics (6)
    • Shrubs and herbs (33)
      • Native (13)
      • Exotics (20)
  • Total number of wetland species (20)
    • Trees (10)
      • Natives (8)
      • Exotics (2)
    • Shrubs and herbs (10)
      • Natives (5)
      • Exotics (5)

 
Comparison of Basal Areas
 
Structurally, the lower section differed from the middle and upper sections. The lower section had the smallest mean diameter at breast height (DBH) (9.5 cm), the highest mean stem density (1585 stems/ha), and the lowest mean basal area (20.1 m2/ha). In contrast, the middle and upper sections had similar means for DBH (11.6 and 10.6 cm), total density (1235 and 1288 stems/ha), and basal area (27.7 and 25.2 m2/ha), respectively.
 
Differences in composition throughout the watershed are shown by a comparison of basal area of the most important species. Among native species, Fraxinus pennsylvanica, Acer negundo, Acer saccharinum, Liriodendron tulipifera and Robinia pseudo-acacia are restricted almost entirely to the middle and lower watersheds, while Acer rubrum and Juglans nigra occur predominantly in the upper and middle sections. Quercus bicolor, Quercus palustris, Acer saccharum, Salix nigra and Platanus occidentalis are predominantly in the upper watershed.
 


 
Comparison of basal areas
 

 
Size distributions of trees and percent cover of herbs show that species have discrete distributions throughout the watershed with small trees of some species occurring in the upper and middle sections and large trees occurring in the lower watershed, and vice versa . Below are plotted histograms of Acer negundo (a typically riparian species) and Acer rubrum (found both in riparian and upland areas).
 

 

 

 

 

 

 
 
 

 

 

 

 

 

 
 
 

 
Non-North American exotic trees are concentrated in the lower watershed. Of the exotic species only 3 (below) are distributed throughout the watershed. The others have very discrete distributions but are abundant where they occur.
 

 

 

 

 
Species distributions with regard to elevations are shown below.. Land surfaces <1 meter above the stream channel elevation support mainly species of trees that normally occupy dry habitats. Red maple is the most abundant species at <1 meter elevation, while box elder is most abundant at elevations >1 meter. The majority of wetland species occurs as large trees on surfaces where elevations are >1 meter, suggesting that these areas represent former floodplains that no longer function as riparian areas. The distribution of species appears to be related to superimposed patterns of stream flooding, bank aggradation and stream incision related to the long history of variable land use as outlined in Table 1. The heterogeneity of past land use is reflected in a heterogeneous riparian area. The occurrence of mainly upland species at the lower elevations suggest a dry conditions related to hydrologic rather than climate change, which is interpreted as a “hydrologic drought”.
 

 

 
 
The Oregon Ridge Vegetation
  • Grace S. Brush, Johns Hopkins University
  • Rachel Myirski, 2008 REU student, Johns Hopkins University
Following up on the results from the riparian vegetation analysis, which indicated that urban land use has resulted in a “hydrological drought” in which riparian trees are being replaced by upland species because of inaccessibity of water due to lowering of the ground water table and changes in the flooding regime, it was decided to investigate the distributions of tree species in Oregon Ridge State Park. The park spans a 1,043 acre (4.22 km2) area in Cockeysville, Maryland and is adjacent to the northernmost part of the Gwynns Falls watershed. Thirty plots were located along a riparian zone in the park and sampling was done as previously described.
 

 

 
Several tree ring cores were obtained from Liriodendron tulipifera and Fagus grandifolia, the two most important species in the park in order to analyze the history of tree growth and water availability. These data were compared with climate records.
 
History of Oregon Ridge
 
Although Oregon Ridge is largely forested today, it has a long history of different land uses. In the mid 19th century, an iron ore and marble mining operation took place within the park, where an iron smelting furnace was built along the Oregon Branch stream. The furnace was in operation from 1844 until 1858. As a result of these operations, an industrial village developed which remained until the 1930s. In addition to the large scale mining and smelting operations, farmlands surrounded the region. Agriculture included both crop production and grazing. Eventually farming was abandonedDuring this time the park was mostly deforested. The trees that are presently there are no more than 70 to 80 years in age.
 
Tree Data
 
Size
Common Name Scientific Name <10 cm 10-19 cm 20-29 cm 30-39 cm 40-49 cm 50+ cm Total
Alternate Leaf Dogwood Cornus alternifolia 12          12
Bald Cypress Taxodium distichum    1 1    2
Beech Fagus grandifolia 46 1        47
Big Toothed Aspen Populus grandidentata 7          7
Birch Betula sp. 2          2
Bitternut Hickory Carya cordiformis 3          3
Black Birch Betula lenta 6          6
Black Gum (Sour Gum) Nyssa sylvatica 54 6 1      61
Cherry Prunus sp. 2          2
Chestnut Oak Quercus prinus    3  2 1 6
Flowering Dogwood Cornus florida 3          3
Green Ash Fraxinus pennsylvanica 1          1
Mockernut Hickory Carya tomentosa 6  1      7
Mossy Cup Oak Quercus macrocarpa        1 2 3
Mountain Laurel Kalmia latifolia 6          6
Mountain Maple Acer spicatum  1        1
Pale Hickory Carya pallida 5 1 2 1    9
Paw-Paw Asimina triloba 23 5  1    29
Red Maple Acer rubrum 6 7 1 2  1 17
Shagbark Hickory Carya ovata 2 1  2    5
Silver Maple Acer saccharinum 1  1      2
Slippery Elm Ulmus rubra 56 4 1      61
Smooth Alder Alnus serrulata 2          2
Tulip Poplar Liriodendron tulipifera  4 1 4 10 14 33
Unknown n/a 11          11
Water Elm Planera aquatica 3          3
White Oak Quercus alba    1 1 2 3 7
Winged Elm Ulmus alata 22 2        24
Witch Hazel Hammamelis virginiana 277 1 1      279
Yellow Birch Betula lutea 6          6
Sizes and numbers of all trees from 30 plots

 
Using these data, histograms of the size distributions of the more common species show graphically the growth patterns of trees in the park.
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The tree ring part of this study focused on tulip poplar (Liriodendron tulipifera) and beech (Fagus grandifolia), the histograms for which are shown below.
 
 

 
Sizes of Lirodendron tulipifera.
 

 

Sizes of Fagus grandifolia.
 

 
By measuring the distance between the rings, the age of each tree and the amount of growth in individual years could be determined.
 

 

 

 

 
Tulip poplar and beech were analyzed with respect to the pattern of succession in the park, since these are the two most numerous trees in the study area. With respect to tulip poplar, no tulip poplar saplings were found in any of the 30 plots sampled. All of the tulip poplar trees are in the larger diameter classes, and presumably older. On the other hand, most of the beech trees were in the sapling category with very few larger (older) trees. This would seem to indicate that tulip poplar is being replaced by beech. Analysis of the tree rings would indicate that hydrologic change is not the reason for the difference between the two species because the growh rates of both species appear similar. However, unusually dry years occurred between 1995 and 2000, during which time tulip poplar growth rings were smaller than before and after this period, whereas beech growth rates are greater during this period. The results, though preliminary, indicate that these species are sensitive to dry and wet climate periods, but whether they are responding to changes in hydrology is not evident.