WEE- WATER AND ENVIRONMENTAL ENGINEERING
Department of hydraulics and hydrology, Faculty of Civil Engineering,
CTU Prague
Problem 1
Do an annual water balance analysis on the watershed adjacent to the gauge station SAN JOSE CREEK
NEAR GOLETA (CA) USGS code: 11120500 with adjacent drainage area 5.51 square miles (14.27 km2).
Details at the nearest climatic station:
SANTA BARBARA MUNICIPAL AIRPORT, CA US; GHCND:USW00023190
GPS: 34.4258°, -119.8425°
Global Summary of the Year: http://www.ncdc.noaa.gov/cdo-
web/datasets/GSOY/stations/GHCND:USW00023190/detail
Gauge station details:
SAN JOSE CREEK NEAR GOLETA (CA) USGS code: 11120500
GPS: 34.4592°, -119.8091°
http://waterdata.usgs.gov/nwis/inventory?agency_code=USGS&site_no=11120500
Picture of watershed of the SAN JOSE CREEK with position of gauging station USGS No:11120500
Indicative information about pan evaporation can be obtained from the nearest climatic station
outside the analysed watershed
CACHUMA LAKE, CA US; COOP:041253
GPS: 34.5822°, -119.9817°
Global Summary of the Month: http://www.ncdc.noaa.gov/cdo-
web/datasets/GSOM/stations/GHCND:USW00023190/detail
Use record of discharge observation during period coincident with period of operation of the nearest
climatic station which provides data on precipitation based on global summary of the month. Try to
carry out an annual water balance during each individual year separately (use L’vovich’s water
balance).
A. Elaborate following table for annual L’vovich’s water balance during at least 2-year period
(1939+2A;1940+2A) where A is student’s number. For student’s number A >37 take the year
1903+A and the year 2053-A, (if data in some year is missing, take data of the following
year). Compare all members of the annual water balance for each single year.
B. Then provide analysis for all observed peak discharges in station SAN JOSE CREEK NEAR
GOLETA (in period 1941-2015) and then in the timeframe for a set of n=30 peak discharges
starting from year 1941+A. If the year 1999 (missing data) is contained in your set of 30
years, extend the timeframe to keep a set of 30 data required. Determine the peak discharge
values for a range of probability of exceedance (0.1, 1, 5, 10, 20, 50, 80, 90, 95, 99, 99.9%) .
Try to estimate it from empirical and theoretical cumulative frequency distribution curve
assuming that probability distribution function Log-Pearson III suits the best. Compare both
estimates of peak flows (for long and for 30-year period) and make a comment on results.
Solution
A. A methodology for catchment water balance has been formulated based on L'vovich's
equation set (Eqs. 5-10). The methodology use all members in annual water balance in
[mm/year] units . The step-by-step procedure follows.
1. Assemble an array of annual precipitation data P (mm/year).
2. From this link to annual statistics (based on calendar year 1) at the gauge station take a
value in an appropriate year (conversion from cf/s to m3/s) and compute runoff height R
from eq. 𝑅𝑖 = 𝑄𝑚𝑒𝑎𝑛 𝑖 × 365 × 86.400/𝐴𝑟𝑒𝑎𝐵𝑎𝑠𝑖𝑛 , where Qmean i [m3/s] , AreaBasin [km2], Ri
[mm/year].
3. Assemble the corresponding discharge hydrograph (at the gauge station) → from this link to
daily data at the gauge station download all data in required calendar year (specify begin and
end date and suitable output format). Set of 365 (366) Q daily data plot in hydrograph. Using a
suitable baseflow separation technique (see picture bellow), separate the discharge
hydrographs into surface runoff S (m3) and U (m3) – eg. in EXCEL use function Qbaseflow i =
MIN{Qi, …, ..Qi+param} , where param ~ 10 – 15
4. Divide S and U by catchment area (A) to obtain S (mm) and U (mm) (in similar way as in point
2 above).
5. With S and U, use Eq. 10 to calculate runoff R (mm).
6. With P and S, use Eq. 6 to calculate wetting W (mm).
7. With U and W, use Eq. 7 to calculate vaporization V (mm).
8. With U and W, use Eq. 11 to calculate the baseflow coefficient Ku.
9. With P and R, use Eq. 12 to calculate the runoff coefficient Kr.
Average values of P, R, S, U, W, and V can be used as overall indicators of the catchment
water balance.
*/ It is possible to perform above analysis also here (http://ponce.sdsu.edu/onlinewaterbalance.php).
1
Note: If there isn´t probable the formation of snow cover, it is possible to carry out each annual water balance
within a calendar year (from January 1 to December 31). If the formation of snow cover is probable, is better
to carry out annual water balance based on water year which is scheduled on the north hemisphere from
October 1 of the preceding year to September 30 of the current year.
Baseflow separation method
B. Peak discharge analysis
1. Rearrange discharges Qpeak in descending order - order of the individual value is m (m is
ranging from 1 to 30 or 74). The number of values is n (here n=30 or 74).
2.
3. Determine probability p of occurrence according to
4. Plot relationship between Qpeak and p in a chart of empirical exceedance curve (cumulative
frequency curve) - use probability paper attached. Magnitude of Qpeak values is on the
ordinate and the corresponding plotting position on the abscisa, representing the probability
of exceedance. Take from this chart quantiles with probabilities 5, 50 and 95 % denoted Q5,
Q50 and Q95 , respectively.
5. Determine parameter s (skewness)
6. From table 1 take for parameter s corresponding skew (asymmetry) coefficient of Cs.
7. From the same table 1 read the values Φ50, and Φ5 - Φ95.
8. Compute value of standard deviation σ, mean value and coefficient of variation Cv
according to the following formulae (methods of quantiles)
You can also compute above statistics with use common formulae for method of moments
(see table and formulae bellow)
mean value
variance
coefficient of variation
skewness (3rd moment about the mean)
skew (asymmetry) coefficient
9. Compute values of theoretical cumulative frequency curve for given probabilities pi
according to
where values Φi read from the table 2 for the Cs. (For Cs<0 and probability pi take inverse
value for -Cs and probability (100- pi) from table)!
10. Plot the theoretical cumulative frequency curve (relationship of Qi to pi) at the same chart as
the empirical curve and compare them each other. Do this for all time frames. Provide all
statistical values ( , σ, Cv, Cs) and estimate of Qi from theoretical c.f.c. for 100 years and
1 000 years in table.
11. Provide and compare range of Qpeak values (in [m3s-1] units) obtained from both curves in the
same confidence interval (0.1%; 99.9%) in all time frames. Explain why is use of the
theoretical line preferred?
**/ It is possible to check peak flow analysis here http://ponce.sdsu.edu/onlinepearson2.php or with
use of a suitable application software eg. http://www.mathwave.com/
-------------------------------------------------------------------------------------------------
!!! Send solution of the problem to address:
[email protected] with e-mail subject “WEE – problem 1, A= {provide your student’s number}”
by the 1st of November.
TABLE 1: Determination of statistics by method of quantiles
TABLE 2: Frequency factor F(Cs,pi) after Foster-Rybkin
(Note: for negative Cs invert the table, i.e. F(-Cs, p [%])=-F(Cs, 100-p [%]))
