Water Balance in Catchments
275
climates it can be an even higher proportion of the total EA (Gash and
Stewart 1977). Thus, in Mediterranean climates, the vegetation controls to a
high degree the hydrological cycle (see also Chaps. 10-12 and 15).
To correctly understand the relationship between the main components of
the water cycle at the catchment level, the potential evapotranspiration (ET)
also has to be considered. In fact, the key variable to explain the water partition under different climates is the relationship between the available water
and the energy available to evaporate that water. This relationship can be
roughly estimated by the ratio P/ET. This ratio is approximately 0.5 at Prades,
1 at Montseny, and greater than 2 in cool-temperate humid areas like the UK
or northeastern USA.
19.3 Interannual Variability of Streamflow
and Evapotranspiration
The main factor that introduces variability in the annual water balance at our
holm oak sites is the variability in annual precipitation. In comparison with
it, annual variations of ET are much lower (Pinol et al. 1998). In rainy years
more water is available both to be evaporated and to be lost by streamflow, so
we could expect a positive relationship between annual P and annual Q, and
between annual P and annual EA. However, as we will see, the interannual
variation of Q and EA depends strongly on climatic characteristics, summarized in the ratio P/ET, in addition to their evident dependence on P.
At Prades, the relationship EA-P is highly significant, whereas the relationship Q-P is only slightly significant for the Avic and Teula catchments (Table
19.2). At Montseny, both relationships are statistically significant, but the first
one (EA-P) is less significant and the second one (Q-P) more significant than
at Prades (Table 19.2). In wetter catchments, like those of Hubbard Brook in
northeastern USA, Q is linearly dependent on P, but EA and Pare uncorrelated (Likens et al. 1977). If the same relationships are derived for different
catchments around the world (Table 19.2), there is a continuum of situations
between those observed at Prades and Hubbard Brook. The slopes for linear
regressions of Q on P range between 0 (Q is not dependent on P) and 1 (all
the Q variation can be attributed to P). Table 19.2 shows the regressions for a
selected group of catchments arranged according to values for regression
slopes of Q on P, and ordered from Mediterranean to humid temperate
catchments. The Montseny catchments occupy an intermediate position between the two groups.
275
climates it can be an even higher proportion of the total EA (Gash and
Stewart 1977). Thus, in Mediterranean climates, the vegetation controls to a
high degree the hydrological cycle (see also Chaps. 10-12 and 15).
To correctly understand the relationship between the main components of
the water cycle at the catchment level, the potential evapotranspiration (ET)
also has to be considered. In fact, the key variable to explain the water partition under different climates is the relationship between the available water
and the energy available to evaporate that water. This relationship can be
roughly estimated by the ratio P/ET. This ratio is approximately 0.5 at Prades,
1 at Montseny, and greater than 2 in cool-temperate humid areas like the UK
or northeastern USA.
19.3 Interannual Variability of Streamflow
and Evapotranspiration
The main factor that introduces variability in the annual water balance at our
holm oak sites is the variability in annual precipitation. In comparison with
it, annual variations of ET are much lower (Pinol et al. 1998). In rainy years
more water is available both to be evaporated and to be lost by streamflow, so
we could expect a positive relationship between annual P and annual Q, and
between annual P and annual EA. However, as we will see, the interannual
variation of Q and EA depends strongly on climatic characteristics, summarized in the ratio P/ET, in addition to their evident dependence on P.
At Prades, the relationship EA-P is highly significant, whereas the relationship Q-P is only slightly significant for the Avic and Teula catchments (Table
19.2). At Montseny, both relationships are statistically significant, but the first
one (EA-P) is less significant and the second one (Q-P) more significant than
at Prades (Table 19.2). In wetter catchments, like those of Hubbard Brook in
northeastern USA, Q is linearly dependent on P, but EA and Pare uncorrelated (Likens et al. 1977). If the same relationships are derived for different
catchments around the world (Table 19.2), there is a continuum of situations
between those observed at Prades and Hubbard Brook. The slopes for linear
regressions of Q on P range between 0 (Q is not dependent on P) and 1 (all
the Q variation can be attributed to P). Table 19.2 shows the regressions for a
selected group of catchments arranged according to values for regression
slopes of Q on P, and ordered from Mediterranean to humid temperate
catchments. The Montseny catchments occupy an intermediate position between the two groups.