Local-scale measurements and modeling
Local-scale measurements and modeling demonstrate that relatively heavy summer precipitation in Arizona generally begins within several days after northern Gulf of California (GC) sea surface temperatures (SSTs) exceed 29°C. The mechanism for this relates to the marine boundary layer (MBL) over the northern GC. For SSTs < 29°C, GC air is capped by a strong inversion ~ 50-200 m above the surface, restricting GC moisture to this MBL. Note SSTs ≥ 29.5°C (i.e. SSTs ≥ threshold SST) correspond with inversion caps 55% in lowest 2 km (this was generally true for higher levels too). So, the inversion generally disappears once SSTs exceed 29°C, allowing MBL moisture to mix with free troposphere. This results in a deep, moist layer that can be advected inland to produce thunderstorms. - More details can be found in the below figures and also in our Journal of Geophysical Research (JGR) paper: Erfani and Mitchell (2014), and in our Atmospheric System Research (ASR) meeting poster: Erfani et al. (2013).
GC = Gulf of California | SSTs = Sea surface temperatures | MBL = Marine boundary layer | R/V = research vessel | RH = relative humidity | OD = optical depth |
Analysis of the 2004 Monsoon
The local-scale analysis is based on soundings lauched from a research vessel (R/V) in the Gulf of California (GC) during June and August, 2004, during the North American Monsoon Experiment (NAME). The figure below shows the location of R/V soundings (blue circles) and approximate R/V cruise path (blue lines) in GC (a) during June 2004 and (b) during August 2004 (Erfani and Mitchell, 2014).
The figure below represents a vertical profile of (a) temperature and (b) relative humidity (RH) for 10% of R/V rawinsondes having the strongest inversion cap; and (c) temperature and (d) RH for 10% of R/V rawinsondes having the weakest inversion cap. The black solid lines show the mean profiles; grey shaded areas represent the range of 1 standard deviation from the mean profiles; black dashed lines depict the adiabatic lapse rate. (a) And (b) were associated with the lowest SSTs (~25 °C), whereas (c) and (d) corresponded to the highest SSTs (~ 30 °C) (Erfani and Mitchell, 2014).
The figure below shows idealized soundings summarizing an MM5 modeling study showing the dependence of the MBL inversion on SSTs when the GC SST = 29°C (green) and when the GC SST = 30°C (red dashed)
The figure below shows (a) Dependence of inversion cap on mean SSTs and (b) dependence of low-level RHs on inversion cap based on mean low-level RH for all rawinsondes on board the R/V in the GC north of 24.1°N during June and August 2004. Bars are indicative of standard deviations, and the numbers on data points represent the frequency of data in their bins. Data points with no standard deviation are based on less than three rawinsondes (Erfani and Mitchell, 2014).
Analysis of the 2012 Monsoon
Represented below is 24 hour rainfall amounts for the periods 10 days before (left) and after (right) the SST threshold in the lower 2/3 of the GC was attained.
Below is shown as (a) Relation between SSTs (red) and convective cloud frequency and (b) relation between convective cloud frequency and cloud top height over the NAM core region during 2012 before (red) and after (blue) the SSTs first exceeded 29°C in the lower 2/3 of the GC. Regarding (a), uncertainties were estimated from the difference in mean SSTs for the central and southern GC regions. OD means cloud optical depth.