Water Technology and Sciences - July - August, 2014 - page 92

90
Water
Technology and Sciences.
Vol. V, No. 4, July-August, 2014
López-de la Cruz and Francés,
Low-Frequency Climate Variability in the Non-Stationary Modeling of Flood Regimes in the Sinaloa and...
Figure 4. Summary of the results obtained using the external covariables model at six representative stations. The results
show the calculated quantiles for different non-exceedance probabilities (0.025, 0.25, 0.50, 0.75, 0.95).
graphs for the
q-q plot
quantiles without trends,
obtained from the review of the normality of
the residuals.
Returning to the inspection of the results in
Figure 4, the calculation of the middle quantile
(with an exceedance probability of 0.50) tended
to be less affected by climate variability.
Nevertheless, the effects are significant in the
calculation for the higher quantiles. Another
point to be noted is that in spite of an adequate
fit of the models, some noise in the results of the
model can be observed, which reflects the non-
linearity in the response of the flood regimes to
natural climate-forced variability.
Comparison Between Non-Stationary and
Stationary Models
The main objective of the study of floods in
operational hydrology is to calculate growth
events for an exceedance probability defined
a priori
in order to obtain flood maps, design
protection measures or establish flood risk
management plans. In fact, in Mexico and
many parts of the world, legislation related
to the risk of floods is based on the analysis
of the frequency of floods to calculate design
floods associated with different time periods
(for example, 20, 50 and 100 years ). These
return periods are related to the need for sound
structures.
Figure 6 shows the results from the
analysis of the frequency of floods in stationary
conditions and non-stationarity conditions for
an exceedance probability of 0.01 (that is, a
return period of 100 years). The graphs show
the problem with assuming stationarity in
the calculation of the flood events. As can be
observed, the non-stationary models indicate
the existence of important periods duringwhich
the flood calculated under non-stationarity is
above that obtained with the stationary model.
Focusing on the analysis of the results for
station 10027, based on the trend model, the
flood for an annual exceedance probability of
0.01 for 58 years of records has ranged from
a minimum value of 640.40 m
3
/s in 1996 to a
maximum of 1 730.52 m
3
/s in 1966. Based on
the covariables model, the minimum value was
295.51 m
3
/s in 1988 and the maximum was 3
508.08 m
3
/s in 1969.
Analyzing the results from the modeling of
the flood events in the non-stationary scenario
with the covariables model and the stationary
1...,82,83,84,85,86,87,88,89,90,91 93,94,95,96,97,98,99,100,101,102,...200
Powered by FlippingBook