Analysis of Streamflow Trend in the Susquehanna River Basin, USA - Hydrologic Time Series Analysis: Theory and Practice - page 181

during a relatively long period of time and is expected to continue into the
future. If the trend is abrupt, the change is a level shift that occurs over a
relatively short time. Once the level shift is completed, the streamflows remain
stationary at the new level until another change occurs (McCabe and Wolock,
2002). The hydrologic literature has so far devoted very limited attention to
the characterization of trend patterns. Kalra et al. (2008) studied the step
change in 639 U.S. streamflows over period of 1951-2002 using the Rank
Sum test and Student's t -test. However, in their study, a prior hypothesis of a
time of change was required for detecting the step change. Miller and Piechota
(2008) investigated the step change in monthly hydroclimatic variables around
the Colorado River Basin. They divided the time series into two sub-series for
the step change analysis which requires the prior knowledge of the time of
change. McCabe and Wolock (2002) employed multiple nonparametric
statistical trend tests with various record lengths to detect change of streamflow
across the conterminous United States. Changes in streamflow could be detected
by examining the number of stream sites with significant trends. To employ
the method suggested by McCabe and Wolock (2002), a group of stream
gauges are required. Hydrology literature is often focussed on large spatial
scale (national or regional) flow change analyses with a large number of
stream gauges. Zhang et al. (2010b) found that their coarse spatial resolution
limits the practical application to specific watersheds in the SRB. To detect a
change in streamflow for single streams, Zhang et al. (2010a) proposed to use
the nonparametric trend test with varying record lengths. Zhang et al. (2010b)
applied this approach as a screening tool to identify potential changes in
streamflow of watersheds within the SRB.
This study expands on the work of Zhang et al. (2010b). The method
proposed by Zhang et al. (2010a) is employed to detect trend patterns in
streamflow in the SRB. Both annual and monthly streamflow characteristics
are investigated. The two components of streamflow, i.e. 'baseflow' and 'storm
runoff', have been examined. Baseflow and storm runoff are estimated by
employing the baseflow separation programme, BFI which is available at the
U.S. Bureau of Reclamation website (http://www.usbr.gov/pmts/hydraulics_lab/
twahl/bfi/). Three questions have been addressed in this study: (i) Are there
changes in the streamflow in the basin? (ii) If so, what is the pattern of the
change, i.e., are the flow changes increasing or decreasing and are they gradual
or abrupt? (iii) If there are changes, in which month(s) do they occur?
9.2 Study Area
The area of the Susquehanna River Basin (SRB) is 71,250 km 2 , which includes
portions of the states of Pennsylvania, New York and Maryland. It is the
largest tributary of the Chesapeake Bay, comprising 43% of the bay's drainage
area and providing 50% of its freshwater. The length of the Susquehanna
River is about 715 km and it is the largest river within the United States that