Monsoon Modeling

Modeling of the 1999 N.A. monsoon onset; the Las Vegas flood

The following modeling results are from the Ph.D. dissertation of Dr. Dorothea Ivanova (Ivanova, D., 2004: Cirrus clouds parameterization for global climate models (GCMs) and North American monsoon modeling study. Ph.D. dissertation, University of Nevada, Reno, 181 pp.). The Arizona/Great Basin onset of the 1999 North American monsoon was simulated using the MM5 regional scale model. This onset also produced a major flooding event in Las Vegas, Nevada. Resolution of the boundary layer was maximized, based on the Hong-Pan (or MRF) boundary layer scheme.

The results generally reproduce the observations under “local scale mechanism”, with the removal of the GC marine boundary layer inversion as SSTs in the northern Gulf of California (GC) approach 30°C. This increased the water vapor mixing ratio at lower levels over the GC and over the southwestern United States, as low-level winds over the northern GC were from the southeast. This in turn increased thunderstorm activity and rainfall amounts over the Southwest. The northern GC SST-rainfall relationship shown under “local scale mechanism” was reproduced in this modeling study.

MM5 modeling study 2003

The nested grid and courser domain of the MM5 modeling study, conducted in 2003.

model terrain cross section

The model terrain nested grid with lines denoting cross sections 1 (AB) and 2 (CD). These cross-sections were evaluated regarding the evolution of the atmospheric circulation, water vapor mixing ratio and potential temperature during 72-hour simulations. The green square depicts the northern Gulf of California (GC) as defined for our modeling study.

 

4 ocean regions - SST evolution model

The left panel shows the 4 ocean regions (A-D) focused upon in this study, while the right panel shows the observed mean time evolution of sea surface temperatures (SSTs) in these ocean regions, where pre-GOC is region A and N. GOC is region D. Note that the northern GC SST lags behind the SST of other regions until late July, when all regions exhibit an SST ~ 30°C. This SST evolution is modeled in this study, including the impact this has on model soundings over the GC (e.g. inversions), water vapor mixing ratios, winds, convective available potential energy (CAPE) and regional rainfall amounts.

 

MM5 model soundings A-B

MM5 model soundings C

MM5 model soundings over the GC along cross-section AB at 60 h (5 am LST), for 3 conditions. Red curve = air temperature; blue curve = dew point temperature. Sounding A: northern GC SST = 26°C, 28°C further south (regions C, B, and A). Sounding B: northern GC SST = 29°C, 30°C further south. Sounding C: northern GC SST = 30°C, 30°C further south. As northern GC SSTs increase, the inversion over the northern GC decreases and disappears in Sounding C. This same behavior was observed with actual soundings over the GC (see “local scale mechanism”). Note that winds below 500 hPa are southeasterly, approximately parallel to GC axis.

water vapor graphs A-B

water vapor graph C

Water vapor mixing ratios (g/kg) and wind velocity components along cross-section AB at 60 h (5 am LST) for 3 simulations: A, northern GC SST = 26°C, 28°C further south (regions C, B and A); B, northern GC SST = 29°C, 30°C further south; C, northern GC SST = 30°C, 30°C further south. As SSTs along cross-section AB increase, the water vapor mixing ratio at low levels also increases.

water vapor graphs A-B 2

Water vapor mixing ratios (g/kg) and wind velocity components along cross-section CD at 60 h (5 am LST) for the 29°C/30°C simulation (panel A; northern GC SSTs = 29°C, 30°C further south) and for the 30°C/30°C simulation (panel B). 

CAPE graph

Magnitude of convective available potential energy (CAPE) at 60, 66 and 72 h (5 am, 11 am and 5 pm LST; top-to-bottom) for simulations 26°C/28°C, 29°C/30°C and 30°C/30°C (left-to-right). These results are consistent with our observational results under “local scale mechanism”.

CIN graph

Magnitude of convective inhibition energy (CIN) at 60, 66 and 72 h (5 am, 11 am and 5 pm LST; top-to-bottom) for simulations 26°C/28°C, 29°C/30°C and 30°C/30°C (left-to-right). These results are consistent with our observational results under “local scale mechanism”.

 

rainfall amounts graphs

Rainfall amounts (cm) at 72 h (over last 6 hours; 5 pm LST) for the 29°C/30°C simulation (left) and the 30°C/30°C simulation (right). The 26°C/28°C simulation was similar to the 29°C/30°C simulation but with slightly less rainfall.

MM5 model experiments

The table shows all the MM5 model experiments: SST values correspond to the evolution of GC SSTs based on June- August climatology. The plot shows normalized rainfall rates over the Arizona region as a function of the N. GC SST. The observed five year mean values for June-August correspond to the Arizona/New Mexico region defined above. The circles indicate 6-hour means predicted by MM5 at 72 h for AZ, southern Nevada, southern California, and extreme northern Mexico, based on the simulations described in the adjacent table.

These results suggest that the proposed local mechanism applies to most Arizona NAM onsets, but 2013 appears to be an exception, as noted under “Overview”. Nonetheless this mechanism appears to be a major contributor of low-level moisture for the NAM.