Laserfiche WebLink
<br />290 <br /> <br />291 <br /> <br />J. l.AVABRE ET AL. <br /> <br />HYDROlOGICA1. RESPONSE OF A MEDITERRANEAN BASIN <br /> <br />FLOOD ON OCTOBER THE 1st 1990 <br /> <br />,~ <br />... <br /> <br />-~.. <br /> <br />... <br /> <br />t: <br /> <br />-.. <br /> <br />~~ <br />,,' '00 <br />~_._l_""'" <br /> <br />(11$) <br /> <br />'" <br /> <br />~.... <br /> <br />'" <br /> <br />: 420 <br />i <br />is 280 <br /> <br />[~=-] <br /> <br />~.. <br />...~. <br /> <br /> <br />'" <br /> <br />,;' <br />'-.' , " <br /> <br />I I <br />.-- <br /> <br />'" <br /> <br />'" <br />Julian ClI1endv (number 01 dey) <br /> <br />'" <br /> <br />Fig. 7. Comparison between the daily runoff observed and simulated by GR3 in 1990. A zoom orthc two <br />most important peaks is shown to allow beuer assessment of the difference. <br /> <br />to locate the start of the permanent flow on 12 October. For this period the <br />basin had some flood responses that have been lost. <br />From the start of the permanent flow, each rainfall gave rise to a flood <br />event. Table 8 shows the four post-fire flood events with peaks exceeding <br />I m's-I km-2 (three of them exceeding 4m's-1 km-2). It is worth noting that <br />the 10 year return period flow (estimated before the fire by a Jenkinson or <br />three-parameter Gumbel law as 6mJs-1) was exceeded three times during the <br />post-fire year. The most surprising is that they are not associated with excep- <br />tional rainfall events. Compare the values shown in Table 8 with the I year <br />return period of accumulated rainfall: 25 mm in I h, 80 mm in 12 h. (These I <br />year return periods were obtained by a partial-duration frequency analysis <br />taking into account the maximum accumulated I hand 12 h rainfalls of each <br />independent rainfall event exceeding the threshold of 13 mm h -I in the case of <br />the I h rainfall, and 35mmh-1 in the case of the 12h rainfall; Lavabre et aI., <br />1990.) . <br />We can see the imporiance of this change in the flood regime by bearing in <br />mind the fact that only fbur events exceeded 4 mJ s -I km -2 during the 23 year <br />reference period (see Table 9). . <br />The post-fire events are comparable in peak flow size, but the I h, and <br /> <br />:0 12 <br />c <br />~ <br /> <br />~ <br />;; 1200 <br />o <br />c <br />o <br />0: 800 <br /> <br />soo <br /> <br />540 <br /> <br />sao <br /> <br />620 <br /> <br />660 <br /> <br />700 <br /> <br />o <br /> <br /> <br />4 <br /> <br />E a <br />E <br /> <br />d <br /> <br />'6 <br /> <br />20 <br /> <br />2000 <br /> <br />1600 <br /> <br />400 <br /> <br />o <br /> <br />Fig. 8. Rimbaud ftood event on 1 October 1990 at a I min time step. <br /> <br />soo <br /> <br />S40 <br /> <br />580 620 <br />lime (minutes) <br /> <br />660 <br /> <br />700 <br /> <br />especially the 12 h accumulated rainfalls are much lower. To surnmarise we <br />can conclude that after the fire the I year return period rainfall is able to <br />produce a 10 year period peak flow, and it was recorded three times in this <br />year. <br />Figures 8 and 9 show two of t~e post-fire flood events at a I min time step. <br />The runoff response of the basin is now produced immediately 'after the <br />rainfall (any a few minutes later) giving rise to important peaks. which also <br />fall very quickly, but with low water yield. <br /> <br />