FLOOD00113

• County: Statewide
• Title: Changes in the Hydrological Response of a Small Mediterranean Basin a Year After a Wildfire
• Date: 6/8/1992
• Prepared For: Journal of Hydrology
• Prepared By: Elsevier Science Publishers
• Floodplain - Doc Type: Educational/Technical/Reference Information

284

J. LAVABRE ET AL

HYDROLOGICAL RESPONSE OF A MEDITERRANEAN BASIN

a part of the annual reduction of the losses previously estimated. Therefore
only the relative evaluation (32%) can be compared with the previous
estimates.
A second monthly analysis was also carried out. A hydrological model,
GR3, proposed by the CEMAGREF group (Michel, 1983) and usually used
as a reference model in France, was used to reproduce the monthly
chronology of the runoff. It is a lumped conceptual model that has only two
calibrating parameters and which needs the rainfall and the potential evapo-
transpiration as inputs, giving the runoff as output (see Appendix). It was
calibrated over the 23 year reference period and used to simulate the monthly
runoff of 1990. Therefore, we can compare the simulated to the observed
runoff to state the change of hydrological behaviour induced by the fire.
In the calibration period this quite simple model provided very good results
(see Table 4a). The mean monthly residuals are relatively small, the Nash
efficiency is E = 0.74 (Nash and Sutcliffe, 1970), showing that the runoff is
estimated fairly and leading to a global difference over the total balance of the
23 year period of less than 1%.
As for the post-fire year (see Table 4b), the comparison between observed
and simulated runoff shows that the values simulated by GR3 are systematic-
ally underestimated, excepting the summer months, where the model cannot
reproduce the very little flow observed, (see Fig. 6). Therefore, assuming that
the model simulation is a good representation of what would have been the
behaviour of the basin if the fire had not existed, we can state again an increase
of the annual runoff of about 148 mm (35% of AR increase) in the new
circumstances. Again, this difference caooot only be due to the hazard fluctua-
tions (less than 3% probability): the residual standard deviation of the
regression between the observed annual runoff and the simulated annual
runoff in the reference period is 0' ~ = 73 mm (R = 0.97). This difference is
once again the hydrological consequence of the fire. We can try to carry this
conclusion further.
It was suggested before that the stated increase of the annual runoff must
be mainly the consequence of the reduction of evapotranspiration due to the
destruction of the vegetation cover (the soil retention power is also affected,
but dealing with annual or monthly amounts the most important consequence
is the reduction of interception and evapotranspiration). More precisely we
should speak about the reduction of interception and transpiration power
related to the vegetation destruction since the effects of the canopy removal
are thought to increa;se evaporation because of the insolation and wind
increases due to the opening up of the forest floor (Chandler et aI., 1983, pp.
185-189). However, in lour case, we are not strictly able to separate evapora~
tion from transpiration losses.

~
W
..J
'"
0(
....

::;

..
C_ 0(
-M
...-:
'I:' ::E
It.
8-'
c ,
~..... \.J,.,
~l:k:'
='"
t I
~Ir:r:~ 0
~~
C Z

'0
o
'C
~
c
.2
;;
:2
..
"
"
-=
c
o
c
,g
.
'"
e
..
,.,
:s
c
c
e
.5
~
'"
"
'0
8
c
.
E
~
~

.5
..
.
c
"
i!
~
"
e
o
::;

0(

N
..;
I

~
..;

'"

'"

~
'"
I

~
..;

'"

'"

~
I

~

~
'"
I

~
~
d

"
~
d
J

~
~
OJ
..J
'"
~

..
oS
~

c
.5
;;
'3
C
'.
~
"
e
c
C

.5
..
,
:~
"
~
"
c
c
::;

285

N

::;

:!(

o

'"

11
C
"
I~ ~- ~
~iC I I
g g 0::;- 0::;.
S! 2 ~ ~
~ :2 i~.n !C:::n
~g t
:z C

1l
>
j
c
c~ tc
8:; 0
!"":":.5
.lI '" .
u ~:
c-~
c t::: C
~ 0 0
~ ~ E
.- ... -c
~ ,., u
-;:2~
.gE"5
_ 0 C
a e '<;ij

;.-
'"

;..
'"
~t,...)!
I
of
'"
~t,...)!

"
~
o
~
u
-5
.5
::2
"
~
o
8
o
~
~
~

1.:-
8 '"
~H~
~ 1-
~
. ",'
c '"
~ r _ lj
o(~_

0(

=

::;

;;

..

.

~

"

~
N

:z

o

o
N

~

0(

N
I

"
;q