Temperature Structure in the ITF

Observations from WOCE XBT section IX1 and PX2 (Fig. 3) are compared with model fields, focusing on potential temperature (Sections 4.1 and 4.2) and wave dynamics in the ITF region (Section 4.3). Examples of the mapped observed and simulated temperature from the IX1 and PX2 sections are shown in Figs. 4-7.

Off the Australian coast (Fig. 4), the yearly formation of the seasonal thermocline dominates variability in the upper 150 m (upper 200 m in model). The amplitude of the seasonal cycle of surface temperature is

Figure 4: Potential temperature as a function of time and depth along IX1 off the coast of western Australia near 25°S for (a) observations and (b) model (For colour version, see Colour Plate Section).

weaker in the model than in reality, and the waters between the seasonal and permanent thermocline have much lower stratification in the model, possibly indicating a too weak Leeuwin Current (and its associated warm advection of subtropical stratification into the coastal zone). Below this depth (~150m), eddy variability is superimposed on a low-frequency modulation of the depth and thickness of the main thermocline. During ENSO years (1986-1987, 1991-1994, and 1997-1998), the thermocline is shallower (closer to the surface), thicker (vertical extent) and surface temperatures are cooler. Although the model reproduces the observed features, the simulated thermocline is too shallow and the ENSO signals are shallower and much weaker than is observed. In particular, the cooling of SST associated with a shallow thermocline, due to lifting and entrainment of cold

H 400

500 600 700

H 400

500 600 700

1984 1986 1988 1990 1992 1994 1996 1998 2000 Year

(b)

0

100

200

£

300

.c

e

400

Q

500

600

700

1984 1986 1988 1990 1992 1994 1996 1998 2000 Year

Figure 5: Potential temperature as a function of time and depth along IX1 off the coast of Java at the Sunda Strait for (a) observations and (b) model (For colour version, see Colour Plate Section).

subsurface water, observed in the 1997/1998 El Nino, is not generated in the model.

Off Java, the thermocline is very thin and lies near 100 m with a warm isothermal layer above it (Fig. 5a, b). While generally reproducing these features, the model thermocline is too thick and deep. A seasonal thermocline is almost entirely absent in both model and observations, and instead, the warm surface layer is punctuated by intense upwelling events the largest of which occur in late 1987, 1988, 1989, 1991, 1994, and 1997. These events are present in the model but persist for longer period and have a stronger SST signature than in reality (Fig. 5b). Both model and observations show upward phase propagation of variations at intraseasonal and semiannual timescales.

Temperatures below 150 m (below 200 m in model) undergo semiannual variations remotely wind driven by the Wyrtki Jets (Clarke and Liu, 1993),

1984 1986 1988 1990 1992 1994 1996 1998 2000

Year

100 200

e400

500 600 700

1984 1986 1988 1990 1992 1994 1996 1998 2000

Year

500 600 700

500 600 700

1984 1986 1988 1990 1992 1994 1996 1998 2000

Year

Figure 6: Potential temperature as a function of time and depth along PX2 off the Java shelf break near 116°E for (a) observations and (b) model (For colour version, see Colour Plate Section).

which feature upward vertical phase propagation both in the observations and model, but are weaker in the latter (by about a factor of 2). The semiannual variability attenuates toward the surface, where interannual variability becomes stronger. The 10°C isotherm of the model also reveals a global warming trend that might have been caused by the relatively short spin-up time in the model, which did not allow a dynamic adjustment of the deeper layers of the model.

At the western end of PX2 (Fig. 6), at the Java Sea shelf break, a thin shallow thermocline also exists. Variability in the surface layer is weak, lacking the dramatic variations seen off Java and is dominantly annual and semiannual, while semiannual and lower frequency variations occur at depth. Here, the model simulates the mean temperature structure well. The temperature variability in the surface layer is too weak in the model, but the simulated thermocline events are quite well correlated with the

a 400 a

a 400 a

600 700

1984 1986 1988 1990 1992 1994 1996 1998 2000 Year

1984 1986 1988 1990 1992 1994 1996 1998 2000 Year

1984 1986 1988 1990 1992 1994 1996 1998 2000 Year

Figure 7: Potential temperature as a function of time and depth along PX2 at the Arafura shelf break near 133°E for (a) observations and (b) model (For colour version, see Colour Plate Section).

1984 1986 1988 1990 1992 1994 1996 1998 2000 Year

Figure 7: Potential temperature as a function of time and depth along PX2 at the Arafura shelf break near 133°E for (a) observations and (b) model (For colour version, see Colour Plate Section).

observations - note the cooling near 150 m in 1997/1998. Off the Arafura shelf west of Darwin (Fig. 7), a very strong seasonal thermocline forms. Below the seasonal thermocline, the annual period continues to dominate but with striking upward vertical phase propagation, as can be seen in the observations and that is very well simulated by the model.

Was this article helpful?

0 0

Post a comment