Streams and floods

• extremes in hydrology take many forms, from extremes in magnitude and length
• by definition, an extreme event must be rare and unusual
• magnitude and frequency of extreme events can be estimated from ordinary, not as severe events
• floods are very important issues all over the world
• current examples: Venezuela, Mozambique, Hurricane Floyd, Katrina, Thailand
  

The nature and causes of floods (and droughts)

• for a hydrologist, a flood is a discharge rate that exceeds some threshold value, generally floods occur when a river over tops its banks and flows across the floodplain
• the principal causes of floods in the Eastern United States and the Gulf Coast are hurricanes and storms
• the principal causes of floods in the Western United States are snowmelt and rainstorms
  
• in rivers, floods and low flows are expressions of the temporal variability in rainfall or snowmelt interacting with river basin characteristics (basin form, hillslope properties, channel network properties)
• flooding may also be the result of sudden release of water from dams or lakes, ice jams
• land use changes can affect the severity and effects of floods
  
• climate cycles
  
• most techniques described above tend to treat floods or droughts as random events in a stationary series. However, climate and riverflow are clearly non-stationary and follow trends and cycles
• example: The 1922 Colorado Compact apportioned water rights on the basis of the average discharge from 1896-1930 (21 billion m³/y) while in 1931-65, the discharge rate was 16 billion m³/y
• examples for teleconnections: El Nino/ENSO, possibly sunspot cycles, some of these are quite controversial
• global Effects of ENSO: temperature and precipitation
• ENSO and Maize yields in Zimbabwe
  
• current ENSO status (Fig)
  
• long term drought in western US as reconstructed from tree rings (Cook et al., 2004) (Fig)
• floods cause the biggest natural hazard damage in the US, example: Mississippi flood, 1993
  
• damages caused by floods in the US
  
• More than half of all fatalities during floods are auto related, usually the result
  of drivers misjudging the depth of water on a road and the force of moving
  water. A car can float in just a few inches of water.
• definition of a drought even more difficult than the definition of floods
• British Rainfall Organization: absolute drought = 15 consecutive days with less than 0.25mm/day on any day, partial drought: at least 29 consecutive days with a mean rainfall less than 0.25mm/day
• agricultural droughts: 'at least a partial crop failure'
• hydrological drought: actual flow in the rivers is of most concern; length and extremeness of flow below a certain level are important ; e.g.: low flows of 10-day duration that occur no more than 5% of the time
• various drought indices, e.g. the Palmer drought severity index (PDSI)
• PDSI is calculated based on precipitation and temperature data, as well as the local Available Water Content of the soil
  
• NOAA monitoring system
• current situaton (Fig)

The hydrograph

• a graph of river stage or discharge versus time at a point is called a stage or discharge hydrograph (Fig5.1)
• there are about 6000 gaging stations in the US that typically measure stage, which needs to be converted into discharge rate using a calibration or rating curve (Fig5.3, Snake River in Colorado, Q = 76.5*stage^4.1), dimensions!
• a typical hyetograph and hydrograph of a creek in VA (Fig5.4)
• peaks in the hydrograph are called floods, background discharge between peaks is called baseflow
• differences in hydrographs of three streams (Fig5.5)

Flood prediction

• Student Excercise: How can we predict extreme events? Download the daily discharge rate data for the Mississippi at St Louis, MO, (St_Louis_daily_70s.tsv). Convert the file into an EXCEL spreadsheet and plot the timeseries of the discharge. Consider overlaying the data for all 10 water years. Also look at a map that shows the Mississippi River Basin (Fig). Can you see any patterns? When do flood occur and why? How could you predict floods? This Fig shows the discharge rate for 4 years in the 1970s.
• USGS Streamflow Information Program
  

Flood routing

• movement of flood waves, complicated as a result of many factors, which?
• example: Flooding in Central Virginia, June, 1995
• flood warning and flood mitigation depends on how quickly a flood crest travels downstream and how high it gets; example: nested basins of the Potomac river (Fig5.6)
• flood routing: prediction of downstream hydrograph, if the upstream hydrograph is known
• flood routing in rivers and by reservoirs (Fig5.10)
• dV/dt = I-O

Flood frequency analysis

• simplest approach: use worst event on record
• past record key for the future? Statistical techniques use the following approach:
• highest discharges recorded in each year are listed
• the floods are ranked according to magnitude, the largest flood is assigned a rank 1, the second largest rank 2, etc
• The flood statistics are estimated graphically by plotting on normal probability paper the logarithm of discharge for each flood in the annual series against the fraction of floods greater than or equal to that flood; this fraction is given by r/(n+1), where r is the rank of the particular flood (Fig5.13)
  
• r/(n+1) is the exceedence probability for this particular event
• the return period, the average span of time between any flood and one equaling or exceeding it, is calculated as Treturn = 1/(exceedance probability).
• the 100 year flood can then be estimated from the graph
• example: Holiday Creek, VA (Fig5.13)
• another way of estimating flood frequencies (as discussed in precipitation lecture):
• list the maximum annual discharge values (Qmax)
• calculate the logarithm of the discharge rates and make a histogram of the log (Qmax) data
• if log (Qmax) is normally distributed, determine average and standard deviation
• use EXCEL's NORMINV function to determine the the log(Qmax) corresponding to the exceedence probability of interest (e.g. 0.01 for the 100 year flood)
• transform log(Qmax) back to Qmax
• normal distribution works often well with precipitation data, log-normal for discharge
• look at the following two rivers, plot a timeseries of the peak discharge, make a histogram and determine the 100 and 1000 year flood
  
• 07010000 Mississippi River at St. Louis, MO
• 01362500 ESOPUS CREEK AT COLDBROOK NY
• problems: not deterministic, based usually on non-adequate data, climate and terrestrial environment is variable

USGS fact sheet

Global Effects of ENSO: Temperature and Precipitation

Drought indices

National Drought Mitigation Center

NOAA monitoring system

Cook, E.R. et al.: Long-Term Aridity Changes in the Western United States, Science, Vol. 306, No. 5698, pp. 1015-1018, 5 November 2004