Results for the Current Climate

The CYGP has been computed from the 10-year mean large-scale fields. The average over the 3 last decades of the 20-th century (1970-1999) for the different models and for ERA40 are represented in Fig. 1 in order to show the geographical distribution of the cyclogenesis. It can be seen that the different models are able to represent the broad distribution patterns of the area of TC genesis. However differences between models can be noted in the different ocean basins.

Fig. 1 Geographical distribution of the genesis index CYGP. The unit is number of TC genesis per 20 years per 5° x 5°. The model identification is shown above each figure for the other models and ERA40 reanalysis
Fig. 1 (Continued)

Over the South-West Pacific several models extend the cyclogenesis zone eastwards farther east of 150°W than is computed in ERA40 (CNRM_cm3, GISS_ model_e_r, INGV_echam4, INMcm3_0, IPSL_cm4, MPI_echam5, NCAR_ ccsm3_0). This deficiency originates from the fact that these coupled models have the tendency to give a too symmetric distribution of precipitation straddling the equator. This is the well-known problem of the so-called double inter tropical convergence zone (ITCZ) frequently found in coupled models in association with an excessive westward extension of equatorial Pacific cold tongue in SSTs (Mechoso et al. 1995; Dai 2006; Lin 2007a). Some of the models give in addition a too zonal distribution over the North Pacific, and do not reproduce the contrast between the North West and North East Pacific (GISS_model_e_r, INMcm3_0, IPSL_cm4), a problem that could be due in part to the coarse resolution in these models. In several models the cyclogenesis index is underestimated over the North Atlantic (BCCR_bcm2_0, CCMA_cgcm3_1, CNRM_cm3, CSIR0_mk3_0, INVG_echam4, INMcm3_0, MPI_echam5, MRI_cgcm2_3-2a, NCAR_ccsm3_0). About half the models indicate a cyclogenesis area in the South Atlantic near the coasts of Brazil (BCCR_bcm2_0, CCMA_cgcm3_1, CNRM_cm3, CSIR0_mk3_0, GISS_model_ e_r, INGV_echam4, MPI_echam5, NCAR_ccsm3_0, UKM0_Hadgem1). This does not seem to match well with current observations, since only one single case of a hurricane in this area has been ever observed, namely Catarina in March 2004 (Pezza and Simmonds 2005).

All models reproduce correctly the TC genesis area over the South Indian Ocean, and over the Gulf of Bengal in the North Indian Ocean. However several models tend to give too much TC genesis over the Arabian Sea (CNRM_cm3, CSIR-0_mk3_0, GISS-model_e_r, INMcm3_0, IPSL_cm4, NCAR_ccsm3). This may be due to a tendency of these models to overestimate precipitation in this region.

In order to provide results of TC genesis in a more quantitative form, the CYGP has been integrated over the traditional ocean basins in order to provide an estimation of the number of TC generated each year over each basin (Table 2).

The figures from the models are an average over the 3 decades 1970-1999. In order to summarize the model results, the mean of the 15 models with the standard deviation (after the ± sign) has been included. The cyclogenesis computed from ERA40 is shown for comparison, as well as the observed number of cyclones over the period 1973-1992 published in Tsutsui and Kasahara (1996).

The six ocean basins are:

• SWI (South West Indian: 45°S-0°S, 20°E-100°E);

• NWP (North West Pacific: 0°-55°N, 100°E-180°E);

• SWP (South West Pacific: 55°S-0°S, 100°E-220°E);

• NEP (North East Pacific: 0°-55°N, 180°E-American coast);

• ATL (Atlantic: 0-55°N, American coast-360°E).

This table confirms a tendency of many models to overestimate cyclogenesis over the South West Pacific and to underestimate it strongly over the Atlantic, while the cyclogenesis computed from ERA40 is in rather good agreement with the

Table 2 Number of annual cyclogeneses per year simulated in the different ocean basins by the 15 models, ERA40 and observed (Tsutsui & Kasahara 1996)

Model

NI

SWI

NWP

SWP

NEP

ATL

INMcm3_0

3.0

16.5

20.9

20.7

20.0

1.3

GISS_model_e_r

3.2

9.7

21.7

18.6

15.2

6.9

CCMA_cgcm3_1

6.9

9.8

35.4

21.0

8.9

1.1

IPSL_cm4

4.9

14.4

26.6

18.8

16.5

2.5

UKMO_hadcm3

7.5

9.0

24.0

25.0

14.5

3.1

MRI_cgcm2_3_2_a

4.1

10.4

31.9

26.4

10.2

1.7

MIROC3_2_medres

6.5

8.0

24.9

23.0

17.5

1.9

BCCR_bcm2_0

4.4

10.4

23.8

32.6

6.3

3.1

CNRM_cm3

6.2

15.4

23.0

25.5

7.6

1.5

GFDL_cm2_0

3.0

11.3

24.9

21.3

19.8

2.0

CSIR0_mk3_0

4.2

9.0

34.1

23.8

9.8

0.8

MPI_echam5

3.3

9.5

28.0

20.4

14.9

2.5

UKMO_hadgem1

4.1

11.3

25.0

19.3

20.6

2.6

NCAR_ccsm3_0

6.0

15.0

22.5

21.1

14.5

1.1

INGV_echam4

4.1

13.2

24.7

23.8

9.3

4.6

Mean ± Std. Dev.

4.8 ± 1.5

11.5 ± 2.7

26.1 ± 4.4

22.8 ± 3.7

13.7 ± 4.7

2.4 ±

ERA40

4.8

11.4

24.5

18.1

18.7

7.0

Obs. 1973-92

5.2

10.5

26.6

15.6

17.4

8.7

observed values. This is an indication that the opposite bias of some models over the Atlantic and South Pacific may be due to excessive convection over the Pacific simulated in these models. It is also likely that the dependence of the convective precipitation threshold on the mean convective precipitation may enhance the reduction of the convective potential over the Atlantic in the models that generate large convective rainfall over the Pacific. To go further into the explanation of such a low amount of CYGP over the Atlantic basin, we have calculated the average of the three main contributions to the CYGP index: vorticity, shear and thermal potential. Only the thermal potential average may explain the underestimated values of CYGP over the Atlantic for some of the models (not shown). Indeed, the models that show the more pronounced lack of activity are those with the weaker thermal potential. Besides the effect of the convective threshold, which is only effective on the edges of the convective precipitation pattern, one may conclude that the weakness of the CYGP over the Atlantic is mainly due to the lack of convective precipitations simulated by the models over this region. These weak convective precipitations are coherent with cool summer SSTs in the coupled models over the Atlantic (not shown). Several models have quite large cold SST biases over the tropical Atlantic that are associated with dry biases (Dai 2006).

In order to investigate the evolution of the results over time we have displayed in Fig. 2 the evolution of the number of cyclogenesis events computed by means of the CYGP for successive 10-year period as averaged over each of the six ocean basins.

Fig. 2 Time evolution of TC genesis simulated in the six ocean basins by the 15 IPCC models (detailed in legend) and by ERA40 (* symbol). Horizontal axis represents the decades from 1860 to 2000 in the 20C3M simulation and from 2000 to 2100 in the A2 scenario, and vertical axis is the equivalent number of TCs following Gray's definition

Fig. 2 Time evolution of TC genesis simulated in the six ocean basins by the 15 IPCC models (detailed in legend) and by ERA40 (* symbol). Horizontal axis represents the decades from 1860 to 2000 in the 20C3M simulation and from 2000 to 2100 in the A2 scenario, and vertical axis is the equivalent number of TCs following Gray's definition

The 10-year averages illustrate a rather large interdecadal variability, as well as the large differences in the number of cyclogenesis among the different models, as already discussed in Table 2.

Some long-term trends are clearly apparent in most simulations over some ocean basins, even over the 20th century, and they tend to increase over the 21st century. This is the case for the Indian Ocean (NI and SWI) where nearly all models simulate a small increasing trend, and also for the North West Pacific. Over the NEP and SWP the trends are less apparent, and trends over the 21st century are even opposite in different models. Over the Atlantic most models give no trend or a slightly decreasing trend, with the exception of a single outlier. The trends over the 20-th century are not very stable and partly masked by the interdecadal variability. ERA40 gives larger CYGP variability and trends than the models for the last 4 decades of the 20th century. The observations over the past twenty to thirty years (Klotzbach 2006; Kossin et al. 2007, Webster et al. 2005) show only relatively small trends in TC numbers over most ocean basins, with an increasing trend only over the North Atlantic, and a decreasing trend over the Northeast Pacific that are not simulated by most of the models, nor by ERA40. So there appears to be more decadal variability in the model- or ERA40-diagnosed TC genesis than in reality, although with only a few decades of reasonably reliable observations, it is difficult to reach a more quantitative conclusion. In order to analyse these contrasting responses it is useful to examine the distribution of the cyclogenesis at the end of the 21st century.

Was this article helpful?

0 0
Survival Treasure

Survival Treasure

This is a collection of 3 guides all about survival. Within this collection you find the following titles: Outdoor Survival Skills, Survival Basics and The Wilderness Survival Guide.

Get My Free Ebook


Post a comment