Laserfiche WebLink
<br />280 <br /> <br />J. LAVABRE ET AL <br /> <br />HYDROLOGICAL RESPONSE OF A MEDITERRANEAN BASIN <br /> <br />28] <br /> <br />TABLE 3 <br /> <br />Hydrological year <br /> <br />Annual precipitation <br />(mm) <br /> <br />1776 <br />1880 <br />545 <br />625 <br /> <br />Annual runoff <br />(mm) <br /> <br />12]9 <br />1186 <br />II] <br />85 <br /> <br />E 800 <br />g <br />.. 700 <br />~ <br />.. <br />.. <br />0 <br />..J <br />.. 600 <br />~ <br />0 <br />0 <br />.. <br />~ SOO <br />~ <br />E <br />~ <br />.0 <br />~ 400 <br />~ <br />> <br /> 300 <br /> 300 <br /> <br /> <br />Extreme values of AP and AR during the reference period (1967-1989) <br /> <br />.~1990 <br /> <br />1973 <br />]976 <br />1988 <br />1989 <br /> <br />of the 1990 hydrological year: the AP observed, 977 mm, is inside the range <br />of usefulness of the equation. It is slightly lower than the mean, but higher <br />than the two last years which do not exceed 600 mm. <br />However, the estimate of the ann"ual runoff provided by the regression <br />eqn. (2) is 464 mm whereas the observed annual runoff was 576 mm, a difference <br />of 112mm (greater than 2.2 times cr",). Note that accepting a normal distri- <br />bution of the residual, the probability that this difference might be due to a <br />hazard fluctuation is less than I %. Therefore we can conclude that the AR <br />observed in 1990 cannot be explained correctly by the relation obtained <br />during the reference period. This conclusion in fact recognises that the fire has <br />led to significant changes in the hydrological behaviour of the basin. So, we <br />can estimate that the increment of annual runoff due to the change of basin <br />chracteristics, or alternatively the reduction of retention and loss ability of the <br />basin, is about 112mm (24% of supplementary runoff). <br />It could be objected thai this year may have been climatically different to <br />the previous years, and that this may have been the main cause of the <br />non-adequacy of the equation. However, when the annual losses (AP - AR) <br />observed in the Rimbaud basin are compared with those observed in the <br />Vaubarnier basin (not affected by the fire), it is clear that the observation for <br />1990 represents an anomaly in the usual relation between the two basins (see <br />Fig. 3). <br />The 1990 point lies outside the envelope of the pre-fire period AL values. <br />The 1990 Rimbaud AL are extremely low in comparison to the AL recorded <br />in the Vaubarnier, and this difference contrasts sharply with the relatively <br />good concordance between the AL of the two basins during the reference <br />period. In this case the Rimbaud shows the lowest value ever observed <br />whereas the Vaubarn{er shows the highest. <br />Again we can condude that there is a significant reduction of the AL <br />(equivalent to an increase of AR) in the Rimbaud basin due to the destruction <br />of the vegetation cover. As we said before, we can assimilate this loss <br /> <br />500 700 <br />Rlmbaud Annual Losses (mm) <br /> <br />Fig. 3. Comparison of the annual losses (AL) between the Rimbaud (85% destroyed by fire) and the <br />Yaubarnier (not affected by fire) basins for the period 1967-1990. <br /> <br />reduction to the reduction of the real evapotranspiration. Note, however, that <br />using the envelope of Fig. 3, the estimate of the anomaly is much higher than <br />the value estimated by the regression equation (2): about 225 mm, instead of <br />the previous 112mm. It is worth nothing that in 1990 the Vaubarnier basin <br />showed especially high losses due to the effect of the two previous very dry <br />years (even in the case of these small basins, water resources have an interan- <br />nual regulation). Thus we admit that the difference reported in the losses may <br />be overestimated in this case by the special characteristics of the hydrological <br />year and by the fact that the value of225 mm is an upper limit for the estimate <br />of the increment of the AR (64% of increment). <br />Finally it may be interesting to analyse whether the increase of the AR <br />produced by the fire affects the base flow and the quick flow in similar way. <br />Of course this analysis depends on the method used to separate the two <br />generally accepted components of the flow. Nevertheless, we consider this <br />analysis necessary. <br />We used the L'vovich algorithm (L'vovich, 1972), which is the reference <br />separation algorithm proposed by the FRIEND European programme and <br />has been applied in the experimental networks of western Europe (Gustard et <br />at, 1989). It provides a base flow separation using daily flows. We can thus <br />compare the base flow to the total flow. <br />In the 1990 hydrological year, the ratio between base flow and total fiow <br /> <br />