|
GEOCHEMISTRY
OF LOESS ON LONG ISALAND
Vesna Kundić and Gilbert
N. Hanson
Department of
Geosciences
Stony Brook
University
Fig.
1 Map of Long Island showing locations of loess studies
Loess is unconsolidated, wind deposited sediment composed mainly
of silt-sized particles with diameters between 15-50 µm. Loess deposits are homogeneous and show
little or no stratification. Loess
deposits on Long Island range from a few centimeters
to several meters in thickness. The
provenance of Long Island loess in Caumsett State Park
and on the Stony Brook
University campus was previously
studied by Zhong and Hanson (2002) using single mica grain 40Ar/39Ar
ages. Almost all of her mica ages fell
within the age range for micas from the bedrock in Connecticut
and eastern New York. We have
continued the provenance studies in a 2.7 meter loess deposit in a kettle in
Wildwood State Park and also dated the time of deposition of the loess using
OSL (Optically Stimulated Luminescence) and radiocarbon dating (Kundic et al, 2003). The OSL ages were 13,780 ±1,100
years for the bottom of the deposit, 13, 400±1,250 years for the middle and
7,730±690 years for just below the soil line. 14C age of 9,792±55
(calendar age of 11,200 Ka BP) was measured on charcoal grains found below
the 7,730±690 OSL age in the section.
The provenance of loess from Wildwood
State Park was also studied using
single grain 40Ar/39Ar muscovite ages. Muscovite gives
ages mainly between 200 and 400 Ma consistent with ages for mica in loess in
other localities on Long Island Zhong and Hanson (2002) which confirms that the
source of Long Island loess is glacial sediment
derived from the basement rocks to the north in Connecticut
and Massachusetts. The micas
from Wildwood state park have a similar distribution to those from the Stony
Brook University
campus with a slightly greater proportion of 300-400 Ma ages.
  
Fig.2. 40Ar/39Ar
ages of Biotite from Wildwood State
Park and SUNYSB campus. Less from Wildwood SP
has larger proportion of ages in the 300-400 Ma age range.
Fig. 3. Grain size distribution
of loess from Wildwood State
Park is shown as green lines and distribution
from SUNYSB campus shown as blue, yellow, pink and light blue lines. Wildwood
State Park loess is missing the
clay peak and much less sand.
Grain size analysis shows a bimodal distribution with
small sand and a larger silt peak which suggests proximal source. The Stony
Brook campus loess has more sand and clay than the loess from Wildwood
State Park. The higher fraction
of sand and the occurrence of what appears to be eolian
sand below the loess, suggests that the source of the campus loess was even
more proximal Zhong and Hanson (2002). The clay may be a result of
soil processes in that the loess was only one meter thick and the upper 50 cm
which was not studied was clearly affected by soil process.
MAJOR ELEMENTS
Chemical Index of Alteration (Taylor,
McLennan (1985)) for Wildwood
State Park loess is 62. On a
scale where 50 is fresh and 100 is highly weathered
rock a value of 62 suggests mild weathering. The weathering may have occurred
at the source rock, while the material was in transit or at the site of
deposition. Fig. 4 compares the data for loess from Wildwood
State Park (wwk)
from near the top, the middle and bottom of the deposit with average loess
given by Schentger, 1992. Fig. 5 compares the loess
from Wildwood State
Park with the average of loess samples
collected by Vesna Kundic,
from Colorado, the Midwest
and Long Island. Fig 6 compares the data for loess
from Wildwood State
Park with the average of stream sediments in New
England, the likely source of Long Island
loess. Wildwood State
Park loess major element data compare very well
with the New England stream data, except that the Wildwood
Sate Park
loess is depleted in Ca and enriched in phosphate. The depletion in Ca may be
a result of leaching of Ca associated with mild weathering of the loess. It
is not clear why the loess is enriched in phosphate. The phosphate content of
Wildwood loess is highly variable with values that range about the average
for loess from Colorado, the Midwest
and Long Island. Loess is commonly enriched in zircon.
It may be that loess is similarly enriched in apatite.
Fe is
strongly enriched in Wildwood loess compared to average loess from Schentger (1992) but when compared to the average of
loess from the Midwest, Colorado
and Long Island, and the stream sediments from New
England there is no Fe enrichment (Fig. 5). Thus, the difference
in Fe content is probably a characteristic of the different sources for
average loess as compared to Wildwood
State Park loess.
Mn is 2-2.5 times enriched relative to average loess and
shows a fairly large variation about 1 compared to the Midwest
and Colorado loess and the New
England stream sediments. Mn is easily
affected by redox conditions and variation in Mn in the samples may reflect addition and leaching of Mn during mild weathering.
Fig. 4. Wildwood
State Park loess normalized to
average loess from Schentger , 1992. shows large enrichment
in Fe and depletion in Ca. Ca is probably leached from the soil because Long
Island soil is acidic.
Fig 5 Wildwood loess compared with the average of loess
calculated from samples collected in Colorado,
Midwest and Long island. There
is no large Fe enrichment but there is Ca depletion is still present
Fig 6. Long Island
loess normalized to the likely chemistry of the source shows the Ca depletion
and a change in Mn concentration with depth which
may be a consequence of soil processes. P is enriched in comparison with the
potential source.
Our initial hypothesis for the source of the loess at Caumsett State Park,
Stony Brook
University campus and Wildwood
State Park was stream deposits in
the bottom of Glacial Lake Connecticut after the lake drained. To test this
hypothesis we converted radiocarbon ages to calendar ages in New
England and Long Island in order to
compare the OSL ages with the radiocarbon ages.
Table 1 Calendar age and radiocarbon ages
associated with Glacial Lake Connecticut
|
Calendar Age
Ka BP
|
Radiocarbon Age
Ka BP
|
Event
|
|
24 to 22
|
21.8 to 19
|
Retreat of ice from the Harbor Hill moraine {Stone, 1986}
|
|
20
|
17.6
|
Ice sheet at Connecticut
shoreline {Stone, 1986}
|
|
18
|
15.5
|
Glacial Lake Connecticut
drained {Lewis and Stone,1991}
|
|
14.3
|
12.4
|
Seawater covered Glacial Lake Connecticut bottom {Lewis
and Stone, 1991}
|
Table 2 Calendar age and radiocarbon ages
relating to Glacial Lake Hitchcock
bottom and stream sediments as source of loess (Ridge et al 1999).
|
Calendar Age
Ka BP
|
Radiocarbon Age
Ka BP
|
Event
|
|
17.2
|
14.3
|
bottom of the eolian sand
|
|
16.9
|
14.1
|
between lake beds and eolian Sand
|
|
15.8
|
13.1
|
between lake beds and eolian Sand
|
|
15.3
|
12.6
|
lowest peat layer in pingo scar
|
|
14.3
|
12.4
|
base of the lowermost peat body
|
|
14.1
|
12.2
|
age from a small peat bog
|
|
14.0
|
12.1
|
basal peat sample
|
|
13.9
|
11.9
|
lower most section of the peat body
|
|
13.4
|
11.5
|
top of dune
|
Fig.7. Calendar ages in Ka BP, horizontal lines represent
approximate time span of events. Radiocarbon ages have been converted to
calendar ages (Stuiver, et al, 1998). Approximate
position of the front of the continental glacier in calendar years after
retreating from Long Island (LI) {Stone, 1986 #352} when it reached
Connecticut (CT) {Stone, 1986 #352}, Massachusetts (MA) {Uchupi,
2001 #64}, New Hampshire (NH) {Ridge, 1999 #330} and Quebec (QC) ({Ridge,
1999 #330} and reference therein {Miller, 1979 #353}). Light and dark brown
lines represent time after Glacial Lake Hitchcock {Stone, 1992 #333} and
Glacial Lake Connecticut drained {Lewis, 1991 #783} and their lake bottoms were dry with
glacial melt water streams crossing the lake bottoms. Broken blue line represents
the time when fresh water or sea water was occupying the lake basins. The
dark blue line is the time when Long Island may have
had permafrost. The Red line shows period of deposition of loess in Wildwood
State Park kettle hole as
measured with OSL and radiocarbon ages.
DISCUSSION
Most of the 40Ar/39Ar
ages are in the 200-400 Ma age range from both locations Wildwood
State Park and SUNYSB Campus.
Therefore, provenance study shows that the source of loess is bedrock to the north
of Long Island Sound.
Major elements in Wildwood
loess correspond well with the average major elements from North
England stream sediments (USGS NURE 2000NE). Wildwood loess is
lower in Ca which might be result of weathering. CIA (Chemical Index of Alteration)
of 62 corresponds to mild weathering. Phosphorus is highly variable which
might be a consequence of Apatite enrichment or soil processes. Mn also shows large variations which is probably a
consequence of weathering.
Grain size distribution of loess
from Stony Brook and Wildwood State
Park shows some differences. Stony Brook loess
has three peaks; sand, silt and clay, whereas Wildwood loess has sand and
silt but no clay peak. The source of clay in Stony Brook loess could be in
the exposed lake bottom sediments to the north or due to the effects of
weathering since this deposit is thinner and more affected by soil processes
than the Wildwood loess deposit. Furthermore, Stony Brook loess has a larger
sand peak than Wildwood State
Park loess which suggests that its source was
closer. The tunnel valley at Stony Brook is shallow and probably had no ice
in it after deglaciation and could have accumulated
loess soon after deglaciation. If there was
permafrost until about 14,000 years BP on Long Island,
the buried ice in the Wildwood kettle would probably not have melted until
that time. The kettle could not accumulate loess until the ice melted. The
deposition of loess on the campus might have started about 18Ka after Glacial
Lake Connecticut drained and while streams were traveling along the bottom of
the Long Island Sound basin. Deposition of loess in the kettle at Wildwood
State Park may have started later
after Glacial lake Hitchcock drained and the area was no longer affected by
permafrost.
REFERENCES
Lewis RS, Stone JR, (1991) Late
Quaternary Stratigraphy and Depositional History of the Long
Island Sound Basin:
Connecticut New
York, Journal of Coastal Research, SI 11, 1-23,
Fall 1991
Kundic
et al, (2003) http://www.geo.sunysb.edu/lig/Conferences/abstracts-03/kundic/kundic.htm
Ridge JC, Besonen MR, Brochu M, Brown SL, Callahan JW, Cook GJ, Nicholson RS,
Toll NJ, (1999) Varve,
paleomagnetic, and C-14 chronologies for late
Pleistocene events in New Hampshire and Vermont (USA), Geographie
Physique Et Quaternaire, 53 (1): 79-107 1999
Schnetger B, (1992) Chemical-composition of loess
from a local and worldwide view, Neues Jahrbuch Fur Mineralogie-Monatshefte,
(1): 29-47 Jan 1992
Stone JR, Ashley GM, (1992) Ice-wedge casts, pingo scars, and the drainage of Glacial Lake Hitchcock,
Guidebook for fieldtrips in the Connecticut valley region of Massachusetts
and adjacent states, Vol. 2, 1992
Stuiver M, Reimer PJ, Bard E, Beck JW, Burr
GS, Hughen KA, Kromer B, McCormac G, Van der Plicht J, Spurk M., (1998) INTCAL98 radiocarbon age
calibration, 24,000-0 cal BP Radiocarbon, 40 (3): 1041-1083 1998
Taylor SR, McLennan SM, (1985) The continental crust, its composition and evolution: an examination of the geochemical record preserved in sedimentary rocks, Blackwell
Scientific Publications, Oxford (Oxfordshire), 1985
USGS NURE 2000NE http://tin.er.usgs.gov/geochem/doc/groups-cats.htm
Zhong and Hanson (2002), http://www.geo.sunysb.edu/lig/Conferences/abstracts_02/zhong/Zhong-abst-2.htm
|