1. Introduction
Interannual variations in warm-season rainfall, which is closely related to the North American monsoon (NAM), have been an important concern in the Arizona–New Mexico (AZNM) and northwestern Mexico (NWM) regions as warm-season rainfall is important for water resources and other environmental concerns. Hence, understanding and predicting the water cycle associated with the NAM are crucial for long-term hydroclimate forecasts in these regions. Carleton et al. (1990) and Higgins et al. (1998) have shown that wet (dry) monsoons in AZNM tend to follow winters that were wetter (drier) than normal in the Pacific Northwest and drier (wetter) than normal locally. Both studies also suggest that the influence of the previous winter is through SST anomalies in the extratropical Pacific adjacent to the region. Carleton et al. (1990) provided evidence that the impacts of these anomalies on the monsoon are via an increase (decrease) in low-level southerly geostrophic winds over the Gulf of California (GC) and associated northward moisture transport into AZ in years with negative (positive) SST anomalies in the Pacific. Negative (positive) SST anomalies in the Pacific along the coast of Baja California, Mexico, tend to increase (decrease) low-level west–east temperature gradients (land–sea thermal contrasts) and thus low-level winds. In addition to the large-scale circulation variability, regional surface forcing by lands and coastal waters have been suspected as other sources of warm-season rainfall variations in these regions. In a regional climate model (RCM) study, Kim et al. (2000) showed that specifying different vegetation characteristics can alter the simulated warm-season rainfall through local evaporation–rainfall feedback. In another RCM study, Small (2001) show that an extended period of soil moisture anomalies over the Colorado Rockies can affect monsoon rainfall in AZNM and NWM by modifying land–sea temperature contrasts.
One of the most important regional surface elements that may influence the warm season water cycle in AZNM and NWM is the SSTs in the GC. Carleton et al. (1990) found a weak association between wet monsoons in AZ and positive SST anomalies in the one grid point of a 5° × 5° SST dataset that they had included as a part of the GC. The GC has certainly been known as an important source of moisture for monsoon rainfall (Hales 1972; Carleton et al. 1990; Douglas 1995; Maddox et al. 1995; Stensrud et al. 1995; Schmitz and Mullen 1996; Berbery 2001; Anderson and Roads 2001). Wright et al. (2001) analyzed time series of stable water isotopes to support the idea that the GC is an important moisture source for monsoon rainfall in AZNM. Large seasonal and interannual variations in the northern GC (NGC) SSTs (Ripa and Marinone 1989) may affect warm-season water cycle around the GC as evaporation from the GC depends strongly on local SSTs. The observational study by Mitchell et al. (2002) suggested that the NGC SSTs may exert a controlling influence on the onset of monsoon rainfall in AZNM. Specifically, they found that the onset of monsoon rainfall in NWM (AZNM) does not occur until the SSTs in the GC (NGC) exceed 26°C (29°C). In addition, they found that AZNM received two-thirds of its summer rainfall during the periods in which SSTs in the NGC are higher than 29.5°C.
This regional modeling study is designed to investigate the sensitivity of the warm-season rainfall characteristics including the onset timing and amounts of monsoon rainfall in AZNM and NWM, as well as mesoscale circulation around the GC to the NGC SSTs. We present the experimental design and model descriptions below. These are followed by discussions on the results from the control simulations and the impacts of the GC SSTs estimated as the differences between the sensitivity and control runs.
2. Experimental design
Two sets of seasonal simulations were performed to study the impacts of the NGC SSTs on warm-season rainfall in AZNM and NWM and the low-level mesoscale circulation around the GC. Each set of experiments includes six seasonal runs covering the 4-month period June–September of three wet (1984, 1988, and 1990), two dry (1980 and 1993), and one normal (1991) summer in AZNM. The model domain covers the entire United States and Mexico (Fig. 1) with a horizontal resolution of 60 km and irregularly spaced 12 σ-layers in the vertical. At this resolution, 19, 41, and 25 model grid points are placed within the northern, southern, and entrance parts of the GC, respectively. The geographical division of the GC used in this study follows Mitchell et al. (2002) except that the central and southern parts of the GC in their study are combined into an expanded southern region (S in Fig. 1). To resolve the low-level atmospheric fields, seven model layers were placed below the level of σ = 0.7, with the lowest model layer at about 85 m above the surface. Analyses of the model output are performed mainly within the region marked by the smaller box centered on the GC (Fig. 1).
The control and sensitivity runs differ only in the SSTs prescribed within the GC. In the control (CNTL) runs, the GC SSTs are obtained by extrapolating the National Centers for Environmental Prediction (NCEP) Reanalysis 2 (R2) SSTs off the west coast of Baja California as the GC is not resolved in the global model used to produce the R2. The GC SSTs obtained in this way are lower than observed values (e.g., Mitchell et al. 2002) by 5–6 K in the NGC. In the sensitivity (GC6) runs, the SSTs are increased by 2, 3, and 6 K in the entrance, southern, and northern part of the GC, respectively, from those prescribed in the CNTL (Fig. 2). Since this study is focused on the impacts of the SSTs in the NGC, only a small amount of SST increases are implemented in the Southern Gulf of California (SGC) and entrance of the Gulf of California (EGC) regions to prevent sudden SST jumps within the GC. The higher SSTs in the GC6 are similar to observed values in the NGC. All CNTL and GC6 simulations are driven by identical time-dependent large-scale forcing prescribed from the R2 at 6-h intervals through lateral boundary conditions.
The RCM employed in this study, the Mesoscale Atmospheric Simulation (MAS) model (Soong and Kim 1996; Kim and Soong 1996), is a primitive equation limited-area model using the σ-coordinates in the vertical. Dependent variables of the MAS are staggered on the Arakawa-C grid in the horizontal, and on the Lorenz grid in the vertical. The advection equation is solved using the third-order accurate finite difference scheme of Takacs (1985) that is characterized by minimal phase errors and numerical dispersion. A four-class version of the cloud microphysics scheme of Cho et al. (1989) and the Simplified Arakawa–Schubert scheme (Pan and Wu 1995; Hong and Pan 1998) are used to compute the grid-scale and convective precipitation, respectively. The solar and terrestrial radiative transfer is computed using the formulations of Harshvardhan et al. (1987), after the impacts of water- (Stephens 1978) and ice-phase (Starr and Cox 1985) clouds are added. Vertical turbulent mixing is computed using the bulk aerodynamic scheme at the surface, and K-theory within the model atmosphere. The eddy diffusivities within the model PBL are computed by the bulk PBL scheme of Troen and Mahrt (1986), and, in the model layers above the PBL, by the local scheme of Louis et al. (1982) with the asymptotic mixing length of Kim and Mahrt (1992). For more details of the MAS, readers are referred to Soong and Kim (1996).
A four-layer version of the NCEP–Oregon State University–Air Force–Hyrdrologic Research Laboratory (NOAH) land surface model (Kim and Ek 1995) is coupled with the MAS to compute the land surface process. The NOAH predicts the soil moisture content (SMC) and soil temperature (STC) within multiple model soil layers, as well as the canopy water content and snow water equivalence. The temperature and specific humidity at the atmosphere–land interface, along with outgoing longwave radiation and ground heat fluxes, are calculated by solving a nonlinear form of the surface energy balance equation. More details of the NOAH and its verification studies were provided by Mahrt and Pan (1984), Pan and Mahrt (1987), Kim and Ek (1995), and Chang et al. (1999).
The thickness of the model soil layers are 0.1, 0.4, 1.0, and 1.5 m, respectively, from the top to the bottom of the model soil in this study. The initial SMC and STC data are obtained by interpolating the corresponding R2 fields. The STC values are further adjusted by assuming a uniform lapse rate of −6.5 K km−1 to compensate the temperature differences due to the differences in model terrain heights between the R2 and the RCM. The soil texture and green-leaf fraction (GLF) are obtained from Zobler (1986) and Gutman and Iganov (1998), respectively. To be consistent with the assumptions used in the retrieval of the GLF data by Gutman and Iganov (1998), the leaf area index is set to a constant value of 4 in all simulations. The vegetation rooting density within the model soil layers is assumed to be uniform in the vertical.
3. Control simulations
The evolution of the upper air fields around the time of monsoon rainfall onset in AZNM in the CNTL (Figs. 3a,b) agrees well with those in the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis (ER; Figs. 3c,d). Strengthening of the 500-hPa high over AZNM around the time of monsoon rainfall onset, one of the well-known changes in the upper air fields around the time of monsoon rainfall onset in AZNM, in the CNTL (Fig. 3b) compares well with that in the ER (Fig. 3d). The daily geopotential height spatial anomaly correlations between the ER and the CNTL remain above 0.8 and 0.7 for the 300- and 500-hPa levels, respectively, throughout the 4-month period in all six summers simulated in this study (not shown).
Sudden rainfall increases from a dry June to a wet July in AZNM and NWM are well represented in the monthly rainfall in the CNTL as well (Fig. 4). This is another well-known change in the warm-season rainfall in these regions associated with monsoon rainfall onsets (e.g., Higgins et al. 1997; Kim 2002). The CNTL reproduces the rainfall decrease from June to July in the Great Plains that accompanies the rainfall increases in AZNM (not shown). The simulated daily rainfall over AZNM compares reasonably with the NCEP Unified Raingauge Data (URD) daily rainfall analysis (Higgins et al. 2000) as well (Fig. 5). For the years 1984, 1988, 1990, and 1993, the CNTL reproduces the sudden rainfall increase at the time of the observed monsoon rainfall onsets (Higgins et al. 1997, their Table 1) that are marked by the vertical lines in Fig. 5 with good accuracy.
These temporal variations in the simulated upper air fields and rainfall in the CNTL suggest that the CNTL has reproduced important features of the temporal evolutions of the upper air fields and daily rainfall despite the fact that the NGC SSTs prescribed in the CNTL are 5–6 K lower than observations. This result suggests that the temporal variations in the upper air fields and daily rainfall related to the onset of NAM rainfall are mainly determined by seasonal variations in the large-scale circulation, not by the GC SSTs. This is in contrast to the modeling study of Stensrud et al. (1995) but agrees with the modeling study by Mo and Juang (2003).
One of the most notable features in the simulated mesoscale circulation around the GC is the sea-breeze-like diurnal variations in the low-level winds, water vapor fluxes, and rainfall. Figures 6a,b shows the 6-yr mean diurnal variations in rainfall [mm (4 h)-1] around the GC over the 30-day period before and after the onsets of monsoon rainfall in the CNTL. For both periods, rainfall around the GC, especially over the western slope of the Sierra Madre Occidental (SMO), shows diurnal variations with the maximum values in late afternoon and evening (1700–0100 MST), consistent with observations (e.g., Berbery 2001). Nocturnal rainfall is relatively large in the postonset period, especially in AZNM (Fig. 6b). The most notable difference in rainfall between the pre- and postonset period is the intensity of rainfall, not the timing of maximum rainfall.
The low-level mesoscale circulation in the CNTL (Fig. 7) is similar to a sea breeze circulation system between the GC and the lands surrounding it but is extended well into inlands, suggesting that the simulated low-level winds around the GC are affected by both land–sea breeze and mountain valley circulations. Over the western slope of the SMO and southern AZNM, it appears that these two types of thermally driven mesoscale circulations are combined into a single thermally driven regional circulation system (e.g., Douglas 1995). The diurnal variations in rainfall and the low-level winds in Figs. 6 and 7 show that the sea-breeze-like regional circulation plays a crucial role in rainfall around the GC, in accordance with earlier studies (e.g., Douglas 1995; Berbery 2001).
The characteristics of the low-level winds in the postonset period such as the turning of the northwesterlies off the west coast of Baja California into the south-southeasterlies over the GC and the western slope of the SMO, as well as the sea-breeze-like diurnal variations between the GC and surrounding lands are well represented in the CNTL (Fig. 7). Over the eastern Pacific off the west coast of Baja California, the low-level northwesterlies develop stronger (weaker) onshore components during the period from late afternoon to early evening (from late evening to early afternoon). Similar diurnal variations in the low-level winds occur between the GC and the western slopes of the SMO where the simulated low-level winds show clear nocturnal flow reversals, that is, formation of downslope winds, over the western slope of the SMO (0100 and 0500 MST). Also appearing in the simulated low-level winds are strong southerly winds over the NGC. Intensity of this wind component, relative to the winds in the surrounding areas, is especially large from 0100 to 1300 MST, similar to the observed nocturnal low-level jets in the region (Douglas 1995; Berbery 2001). The diurnal variations in the low-level winds during the preonset periods (not shown) are similar to the postonset periods, except that winds, especially those directed from the GC to the SMO and SWAZ, are generally stronger than in the postonset periods. Another important difference in the simulated low-level winds between the pre- and postonset periods is that the south-southeasterly winds over the GC and the low-level jets in the NGC region are much stronger in the postonset period than in the preonset period. These changes in the low-level winds can affect the low-level water vapor fields over the GC as well as the associated moisture fluxes from the GC into surrounding lands as the south-southeasterly winds in the postonset period have longer fetches over the GC that in turn increase moisture supplies from the GC surface to the low atmosphere.
Diurnal variations in the low-level moisture fluxes (Figs. 8a,b) show that the sea-breeze-like circulations between the GC and its surrounding lands, most notably the SMO and southwestern Arizona (SWAZ), and the associated low-level moisture fluxes are important for the simulated rainfall. Figures 8 and 6 clearly show that maximum rainfall in SWAZ and western slope of the SMO occur in the period of maximum moisture fluxes into these regions from the GC. This relationship between the diurnal variations in the mesoscale circulation and rainfall is evident in both pre- and postonset periods, despite some differences in details.
There are several notable differences in the diurnal variations in the mesoscale low-level circulation and rainfall between the pre- and postonset periods. The most notable one is that strong moisture fluxes penetrate further inland over the western slope of the SMO and SWAZ in the postonset period (Fig. 8b) than the preonset period (Fig. 8a). This is thought to be the most direct cause for the overland rainfall increases after the onset of monsoon rainfall in the region. This enhancement in the low-level moisture fluxes in the postonset period must be related to an increase in the atmospheric moisture over the regions upstream side of the SMO and SWAZ, that is, over the eastern Pacific and the GC, as the magnitudes of the onshore winds into the SMO are smaller in the postonset period than the preonset period.
To find the sources of the increased water vapor in the upwind side of the SWAZ and SMO, we examine two likely sources of water vapor, evaporation over the eastern Pacific including the GC (Fig. 9) and the low-level water vapor fields (Fig. 10) in the pre- and postonset periods. Figure 9 shows that evaporation increases over the eastern Pacific and the GC, especially over the SGC and EGC regions, in the postonset periods. The large increases in overland evaporation over the western slope of the SMO and AZNM after monsoon rainfall onset are most likely to be caused by increased rainfall in the postonset period; hence they may not be the cause for the rainfall increases in the postonset period. The large-scale specific humidity over the eastern Pacific off the coast of Baja California and southern California also increases in the postonset period by as much as 1 g kg−1 (Fig. 10a). These increases in the low-level moisture, together with the turning of the low-level winds from northwesterlies over the eastern Pacific into westerlies across Baja California and into south-southeasterlies over the GC (Fig. 7) around the time of monsoon rainfall onset, suggest that the increases in the low-level moisture in the postonset period are associated partially with the increases in the large-scale water vapor over the eastern Pacific, with additional contributions from GC evaporation due to longer fetches and higher SSTs in the postonset period.
4. Impacts of higher NGC SSTs
Rainfall in the GC6 undergoes similar diurnal variations as in the CNTL, but the GC6 runs generate more rainfall around the GC than the CNTL. The largest rainfall increases due to higher GC SSTs occur during the periods of peak rainfall, early afternoon through early night (Figs. 11a,b). In addition, nocturnal rainfall in AZNM also increases substantially with higher GC SSTs in the postonset period (Fig. 11b). The impacts of higher NGC SSTs on the simulated rainfall diurnal variations during the preonset period are similar to the postonset period in their geographical locations and timing; however, the amount of rainfall increases are much larger in the postonset period. These differences in rainfall changes between the pre- and postonset periods in response to higher NGC SSTs in the GC6 are not likely to be solely caused by the differences in the GC SSTs between the pre- and postonset periods. Although the NGC SSTs in the GC6 are lower in the preonset period than in the postonset period, they are similar to or higher than the postonset NGC SSTs in the CNTL (Fig. 2). Still, preonset rainfall in the GC6 (not shown) is much less than postonset rainfall in the CNTL (Fig. 6b), suggesting that the changes in the large-scale circulation and the low-level winds over the GC around the time of monsoon rainfall onset are the primary forcing mechanisms behind the temporal variations in warm-season rainfall in the NWM and AZNM simulated in this study. As discussed in section 3, the low-level south-southeasterlies are more pronounced in the postonset period (Fig. 7). Hence, the low-level winds have longer fetches over the GC in the postonset periods than in the preonset periods. In other words, the low-level winds crossed the GC via much longer paths in the postonset period than in the preonset period. As a result, the low-level air is supplied with more water vapor from the GC in the postonset period than the preonset period, which in turn results in larger rainfall increases in the postonset period. The higher SSTs in the NGC, therefore, may have affected rainfall not by modifying the large-scale circulation, but by enhancing evaporation from the GC into the low atmosphere under given large-scale conditions.
The GC SSTs affect rainfall around the GC through sea-breeze-like mesoscale circulation between the GC and the western slopes of the SMO and SWAZ, as well as the low-level water vapor fluxes associated with it. Differences in the diurnal variations in the low-level water vapor fluxes between the CNTL and the GC6 (Figs. 12a,b) suggest that the rainfall increases in NWM and SWAZ in response to higher GC SSTs are closely related to the increases in water vapor fluxes from the GC. Note that the differences in the low-level water vapor fluxes between the CNTL and the GC6 are directed from the GC into the western slope of the SMO and SWAZ. Also, the regions in which large increases in water vapor fluxes occur (Fig. 12) coincide with those where large rainfall increases occur in both pre- and postonset periods (Fig. 11).
These increases in water vapor fluxes into the western slope of the SMO and SWAZ in response to higher GC SSTs are a consequence of two opposing impacts of higher GC SSTs on regional circulations. The differences in the low-level winds between the CNTL and GC6 in the postonset period (Fig. 13) are directed from the western slope of the SMO and SWAZ to the GC, that is, the onshore component of the low-level winds that are directly related to the magnitudes of moisture fluxes into the land are reduced by higher GC SSTs. This change in the low-level winds, therefore, must have acted to reduce the moisture fluxes from the GC as well as overland rainfall. The reduction in the onshore component of the sea-breeze-like circulation in the GC6 is caused by reduced land–sea temperature contrasts due to lower land surface temperatures in the western slope of the SMO and SWAZ and higher GC SSTs (not shown). The lower surface temperatures in SMO and SWAZ in the GC6 runs are caused by enhanced evaporation and reduced insolation associated with increased rainfall in the region. These changes in land temperatures, in conjunction with higher GC SSTs, result in smaller land–sea temperature differences during the daytime/evening and in turn weaken the onshore winds. Similar changes in the low-level winds in response to higher GC SSTs in the GC6 runs also appear in the preonset period (not shown). These impacts of higher GC SSTs on the low-level winds are countered by increased evaporation from the GC due to higher SSTs in the GC6, especially in the NGC where the largest SST increases are implemented (Fig. 14). Enhanced water vapor fluxes into the land with higher GC SSTs (Fig. 12) suggest that the increases in evaporation from the GC in response to higher SSTs are large enough to overcome the adverse impacts of reduced onshore winds (Fig. 14), resulting in net increases in moisture fluxes and rainfall over the western slope of the SMO and SWAZ.
5. The impacts of higher NGC SSTs on onset timing of monsoon rainfall in AZNM
One of the key questions we attempt to examine in the GC6 runs is whether the warming of the NGC plays an active role in northward migration and monsoon rainfall onset in NWM and AZNM. Although not explicitly mentioned, the modeling and observational studies of Stensrud et al. (1995) and Mitchell et al. (2002), respectively, imply that the NGC SSTs may play an important role in determining the northward migration of monsoon rainfall in the region. Monthly rainfall variations in the GC6 runs (Fig. 15) are similar to those in the CNTL runs (Fig. 4) although the June GC SSTs in the GC6 are higher than the July GC SSTs in the CNTL. With higher GC SSTs in the GC6, rainfall increases generally, especially over the western slope of the SMO. This increase in rainfall in response to higher GC SSTs in the GC6 varies geographically and temporally. In June, generally before the onset of monsoon rainfall in AZNM and NWM, rainfall increases notably over the western slope of the SMO in the vicinity of the SGC and EGC regions where monsoon rainfall onsets are earlier than in the northern regions by almost one month (Higgins et al. 1999, their Fig. 12), while rainfall increases associated with higher NGC SSTs are minimal in AZNM and NWM. Note that the SSTs in the SGC and EGC in the GC6 are higher than those in the CNTL as well. In the July–September period, the months generally after monsoon rainfall onsets in AZNM, rainfall increases over the entire western slope of the SMO and AZNM, with the largest increases in the NWM regions south of the U.S.–Mexico border. Examination of rainfall over the 30-day period before and after monsoon rainfall onset in AZNM (not shown) also shows similar geographical and temporal variations in the rainfall changes in response to higher GC SSTs in the GC6. In addition, June-to-July reduction in rainfall in the Great Plains is not affected much by the higher GC SSTs in the GC6 runs. Hence, the impacts of higher GC SSTs in the GC6 are confined to the region in the vicinity of the GC.
Daily rainfall in AZNM in the CNTL and GC6 (Fig. 16) also show that the higher NGC SSTs alone do not change the temporal variations in rainfall as the CNTL and GC6 generate very similar daily rainfall sequences in AZNM. The higher NGC SSTs show similarly small impacts on daily rainfall in NWM (not shown) as well. This similarity between the daily rainfall sequences in the CNTL and GC6 shows that although higher NGC SSTs in the GC6 runs affect strongly the amount of rainfall and mesoscale circulation characteristics in AZNM and NWM, they have minimal impacts on the temporal variations in, or onset timing of, warm-season rainfall in AZNM and NWM.
6. Conclusions and discussion
We investigate the impacts of the NGC SSTs on warm-season rainfall and mesoscale circulation in AZNM and NWM associated with NAM from two sets of regional simulations. To isolate the differences originating from the difference in the NGC SSTs, the two sets of simulations are run with identical large-scale forcing through lateral boundary conditions except the GC SSTs. In the CNTL runs, the GC SSTs that are obtained by extrapolating the R2 surface temperatures off the west coast of Baja California are lower than observations by 5–6 K. In the GC6 runs, the SSTs are increased by 2, 3, and 6 K in the EGC, SGC, and NGC regions, respectively, from those prescribed in the CNTL.
The CNTL runs simulate well the temporal variations in the large-scale circulation and rainfall that are closely related to monsoon rainfall onsets in AZNM despite the fact that the NGC SSTs are lower than observations by 5–6 K. Strengthening of the 500-hPa high over AZNM around and after monsoon rainfall onsets in AZNM simulated in the CNTL compares well with the ER. Daily spatial anomaly correlation coefficients between the geopotential heights in the CNTL and the ER exceed 0.7 at both the 500- and 300-hPa levels, indicating that the simulations do not drift from the large-scale fields during the simulation periods. The CNTL also simulate well the sudden increase (decrease) in rainfall in AZNM and NWM (Great Plains) from June to July that is one of the notable rainfall changes associated with monsoon rainfall onset in AZNM. Daily rainfall in the CNTL compares reasonably with the NCEP–URD rainfall data, although the onset timing of monsoon rainfall determined from observations (Higgins et al. 1997) is not exactly reproduced in the simulations. The CNTL runs also successfully simulate important features in the low-level winds over the GC region, including the turning of northwesterly winds off the west coast of Baja California into south-southeasterly winds over the GC and the western slope of the SMO, and the low-level jets over the NGC during the postonset period. The sea-breeze-like circulation between the GC and surrounding lands is clearly depicted in the CNTL as well.
Higher GC SSTs prescribed in the GC6 do not have significant impacts on the temporal variations in the large-scale circulation and daily rainfall events in AZNM and NWM. Temporal variations in 500-hPa geopotential heights and monthly rainfall simulated in the GC6 are close to those in the CNTL, except that the rainfall increase after the onset of monsoon rainfall is much larger with higher NGC SSTs. Both the CNTL and GC6 runs generate a very similar sequence of daily rainfall in AZNM over the 4-month period for all 6 yr investigated here despite large differences in the NGC SSTs. The most notable differences in the daily rainfall simulated in the GC6 and the CNTL are rainfall amounts—mostly after monsoon rainfall onset, not the sequence of daily rainfall.
Examination of the temporal variations in the upper air fields and rainfall in the CNTL, as well as comparison of those in the CNTL and GC6, suggests that the northward migration of monsoon rainfall along the GC is determined primarily by the seasonal evolution of the large-scale circulation, not by warming of the NGC. Instead, warming of the NGC may be associated with seasonal variations in the large-scale circulations that drive both northward migration of monsoon rainfall and warming of the NGC. The differences between the CNTL and GC6 runs suggest that the increases in the low-level moisture over the subtropical eastern Pacific off the west coast of southern California and Baja California in conjunction with the low-level winds that develop more pronounced south-southeasterly components in the postonset period may be the primary source of moisture for the monsoon rainfall in AZNM and NWM, with additional contributions from evaporation from the GC. The differences in the low-level winds between the pre- and postonset periods also suggest that the effects of evaporation from the GC on overland rainfall are larger in the postonset period, as the low-level winds have longer fetches over the GC in the postonset period than the preonset period. The longer fetches over the GC during the postonset period can supply more water vapor to the low-level air from the GC surface.
The impacts of the GC SSTs are largely confined to the mesoscale circulation around the GC but play an important role in shaping the water cycle over the land areas around the GC. Both the CNTL and GC6 show that sea-breeze-like diurnal variations in the low-level winds between the GC and the western slope of the SMO play an important role in producing rainfall in the region. In particular, periods of heavy rainfall, from early afternoon to late evening, coincide with the periods of strongest onshore flows, and the associated low-level water vapor fluxes from the GC. Higher GC SSTs in the GC6 tend to reduce the intensity of the sea-breeze-like circulation between the GC and the western slope of the SMO and AZNM compared to the CNTL. This weakening of the sea-breeze-like circulation that should have acted to reduce the low-level water vapor fluxes is countered by the increases in evaporation from the GC. The net result of these two opposing effects of higher GC SSTs is an increase in the low-level water vapor fluxes from the GC that in turn increases rainfall in the western slope of the SMO and AZNM. The impacts of higher GC SSTs are largest in the postonset period, perhaps because of the differences in the low-level winds that tend to have longer fetches over the GC in the postonset period.
The results of this study suggest that the NGC SSTs alone do not affect the onset timing of monsoon rainfall in AZNM and NWM. Although this study shows that accurate GC SSTs are important for quantitative simulations of warm-season rainfall around the GC, it does not support the hypothesis that the evolution of the NGC SSTs can drive the northward migration of monsoon rainfall. Instead, we may hypothesize that there may exist mechanisms related to the June-to-July evolutions of the large-scale atmospheric and oceanic circulations that drive both monsoon rainfall onsets and warming of the NGC, resulting in the close correlations between the NGC SSTs and monsoon rainfall onsets found in recent observational studies.
Acknowledgments
The authors thank W. Higgins and W. Shi at NCEP for making the URD available to this study. This study was supported by NOAA (Grants NA00AANRG0201, NA06GP0376, NA03OAR4310012, and NA17RJ1231) and NASA (Grants NAG5-13248, NAG5-11363, NAG5-11738, and NAG8-1875). The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or NASA.
REFERENCES
Anderson, B. T., and J. O. Roads, 2001: Summertime moisture divergence over the southwestern US and northwestern Mexico. Geophys. Res. Lett., 28 , 1973–1976.
Berbery, E. H., 2001: Mesoscale moisture analysis of the North American monsoon. J. Climate, 14 , 121–137.
Carleton, A. M., D. A. Carpenter, and P. J. Weser, 1990: Mechanisms of interannual variability of the southwest United States summer rainfall maximum. J. Climate, 3 , 999–1015.
Chang, S., D. Hahn, C. Yang, D. Norquist, and M. Ek, 1999: Validation study of the CAPS model land surface scheme using the 1987 Cabauw/PILPS dataset. J. Appl. Meteor., 38 , 405–422.
Cho, H., M. Niewiadomski, and J. Iribarne, 1989: A model of the effect of cumulus clouds on the redistribution and transformation of pollutants. J. Geophys. Res., 94 , D10,. 12895–12910.
Douglas, M. W., 1995: The summertime low-level jet over the Gulf of California. Mon. Wea. Rev., 123 , 2334–2347.
Gutman, G., and A. Iganov, 1998: Derivation of green vegetation fraction from NOANA VHRR for use in numerical weather prediction models. Int. J. Remote Sens., 19 , 1533–1543.
Hales Jr., J. E., 1972: Surges of maritime tropical air northward over the Gulf of California. Mon. Wea. Rev., 100 , 298–306.
Harshvardhan, and Randall, D., and T. Corsetti, 1987: A fast radiation parameterization for atmospheric circulation models. J. Geophys. Res., 92 , 1009–1016.
Higgins, R. W., Y. Yao, and X. Wang, 1997: Influence of the North American monsoon system on the U.S. summer precipitation regime. J. Climate, 10 , 2600–2622.
Higgins, R. W., K. C. Mo, and Y. Yao, 1998: Interannual variability of the U.S. summer precipitation regime with emphasis on the southwestern monsoon. J. Climate, 11 , 2582–2606.
Higgins, R. W., Y. Chen, and A. V. Douglas, 1999: Interannual variability of the North American warm season precipitation regime. J. Climate, 12 , 653–680.
Higgins, R. W., W. Shi, E. Yarosh, and R. Joyce, cited. 2000: Improved United States Precipitation Quality Control and Analysis. Atlas No. 7, NCEP/CPC. [Available online at http://www.cpc.ncep.noaa.gov/research_papers/ncep_cpc_atlas/7/index.html.].
Hong, S., and H. Pan, 1998: Convective trigger function for a mass flux cumulus parameterization scheme. Mon. Wea. Rev., 126 , 2599–2620.
Kim, J., 2002: Precipitation variability associated with the North American Monsoon in the 20th century. Geophys. Res. Lett., 29 .1650, doi:10.1029/2001GL014316.
Kim, J., and L. Mahrt, 1992: Simple formulation of turbulent mixing in a stable free atmosphere and nocturnal boundary layer. Tellus, 44A , 381–394.
Kim, J., and M. Ek, 1995: A simulation of the surface energy budget and soil water content over the Hydrologic Atmospheric Pilot Experiments-Modelisation du Bilan Hydrique forest site. J. Geophys. Res., 100 , D10,. 20845–20854.
Kim, J., and S. Soong, 1996: Simulation of a precipitation event in the western United States. Regional Impacts of Global Climate Change, S. Ghan et al., Eds., Battelle Press, 73–84.
Kim, J., N. Miller, T. Kim, J. D. Farrara, and X. Zeng, 2000: Effects of land-surface characterization on simulating summertime precipitation: Implications on warm-season extended forecasts. Proc. Second Southwest Weather Symp., Tucson, AZ, NOAA/NWS.
Louis, J. F., M. Tiedke, and J. Gelvyn, 1982: A short history of the operational PBL-parametrization at ECMWF. Proc. Workshop on Planetary Boundary Layer Parameterization, Reading, United Kingdom, ECMWF, 59–79. [Available from ECMWF, Shinfield Park, Reading, RG 29AX, United Kingdom.].
Maddox, R., D. McCollum, and K. Howard, 1995: Case study of a severe mesoscale convective system in central Arizona. Wea. Forecasting, 10 , 643–665.
Mahrt, L., and H. Pan, 1984: A two-layer model of soil hydrology. Bound.-Layer Meteor., 29 , 1–20.
Mitchell, D. L., D. Ivanova, R. Rabin, T. J. Brown, and K. Redmond, 2002: Gulf of California sea surface temperatures and the North American monsoon: Mechanistic implications from observations. J. Climate, 15 , 2261–2281.
Mo, K. C., and H. Juang, 2003: Influence of seas surface temperature anomalies in the Gulf of California on North American monsoon rainfall. J. Geophys. Res., 108 .4112, doi:10.1029/2002JD002403.
Pan, H., and L. Mahrt, 1987: Interaction between soil hydrology and boundary layer development. Bound.-Layer Meteor., 38 , 185–202.
Pan, H., and W. Wu, 1995: Implementing a mass flux convection parameterization package for the NCEP medium-range forecast model. NMC Office Note, 40 pp. [Available from NCEP/EMC, 520 Auth Road, Camp Springs, MD 20764.].
Ripa, P., and S. G. Marinone, 1989: Seasonal variability of temperature, salinity, velocity, vorticity and sea level in the central Gulf of California, as inferred from historical data. Quart. J. Roy. Meteor. Soc., 115 , 887–913.
Schmitz, T. J., and S. L. Mullen, 1996: Water vapor transport associated with the summertime North American monsoon as depicted by ECMWF analyses. J. Climate, 9 , 1621–1633.
Small, E. E., 2001: The influence of soil moisture anomalies on variability of the North American monsoon system. Geophys. Res. Lett., 28 , 139–142.
Soong, S., and J. Kim, 1996: Simulation of a heavy precipitation event in California. Climate Change, 32 , 55–77.
Starr, D., and S. Cox, 1985: Cirrus clouds. Part I: A cirrus cloud model. J. Atmos. Sci., 42 , 2663–2681.
Stensrud, D. J., R. L. Gall, S. L. Mullen, and K. W. Howard, 1995: Model climatology of the Mexican monsoon. J. Climate, 8 , 1775–1794.
Stephens, G., 1978: Radiation profiles in extended water clouds. II: Parameterization schemes. J. Atmos. Sci., 35 , 2123–2132.
Takacs, L., 1985: A two-step scheme for the advection equation with minimized dissipation and dispersion error. Mon. Wea. Rev., 113 , 1050–1065.
Troen, I. B., and L. Mahrt, 1986: A simple model of the atmospheric boundary layers; sensitivity to surface evaporation. Bound.-Layer Meteor., 37 , 129–148.
Wright, W. E., A. Long, A. C. Comrie, S. W. Leavitt, T. Cavazos, and C. Eastoe, 2001: Monsoonal moisture sources revealed using temperature, precipitation and precipitation stable isotope timeseries. Geophys. Res. Lett., 28 , 787–790.
Zobler, L., 1986: A world soil file for global climate modeling. NASA Tech. Memo. 87802, 33 pp.
The model domain and the area (the inner box) in which the regional flow fields around the GC are presented.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The model domain and the area (the inner box) in which the regional flow fields around the GC are presented.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The model domain and the area (the inner box) in which the regional flow fields around the GC are presented.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The daily-mean SSTs (°C) in the northern GC from 1 Jun to 30 Sep used in the CNTL (solid line) and GC6 (dotted line) runs. The light (heavy) shading indicates the area in which the geopotential height is less (more) than 5710 (5870) m.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The daily-mean SSTs (°C) in the northern GC from 1 Jun to 30 Sep used in the CNTL (solid line) and GC6 (dotted line) runs. The light (heavy) shading indicates the area in which the geopotential height is less (more) than 5710 (5870) m.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The daily-mean SSTs (°C) in the northern GC from 1 Jun to 30 Sep used in the CNTL (solid line) and GC6 (dotted line) runs. The light (heavy) shading indicates the area in which the geopotential height is less (more) than 5710 (5870) m.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The 500-hPa geopotential heights (m) averaged over the 30-day periods before and after the observed monsoon rainfall onset dates in AZNM for the 6 yr (Higgins et al. 1997), from the (a), (b) CNTL and (c), (d) ER.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The 500-hPa geopotential heights (m) averaged over the 30-day periods before and after the observed monsoon rainfall onset dates in AZNM for the 6 yr (Higgins et al. 1997), from the (a), (b) CNTL and (c), (d) ER.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The 500-hPa geopotential heights (m) averaged over the 30-day periods before and after the observed monsoon rainfall onset dates in AZNM for the 6 yr (Higgins et al. 1997), from the (a), (b) CNTL and (c), (d) ER.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The mean-monthly rainfall (mm day−1) in the CNTL averaged over the 6 yr. The light (heavy) shading indicates the values greater than 1.0 (2.5) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The mean-monthly rainfall (mm day−1) in the CNTL averaged over the 6 yr. The light (heavy) shading indicates the values greater than 1.0 (2.5) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The mean-monthly rainfall (mm day−1) in the CNTL averaged over the 6 yr. The light (heavy) shading indicates the values greater than 1.0 (2.5) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Daily rainfall within the AZNM region in the NCEP–URD data (solid line) and in the CNTL (dashed line) run. The vertical lines indicate the monsoon rainfall onset dates determined by Higgins et al. (1997) from rain gauge analysis data.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Daily rainfall within the AZNM region in the NCEP–URD data (solid line) and in the CNTL (dashed line) run. The vertical lines indicate the monsoon rainfall onset dates determined by Higgins et al. (1997) from rain gauge analysis data.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Daily rainfall within the AZNM region in the NCEP–URD data (solid line) and in the CNTL (dashed line) run. The vertical lines indicate the monsoon rainfall onset dates determined by Higgins et al. (1997) from rain gauge analysis data.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) The 4-h accumulated rainfall [mm (4 h)−1] averaged over the CNTL runs over the 30-day periods before the onsets of monsoon rainfall in AZNM. The monsoon onset dates were adopted from Higgins et al. (1997). The light (heavy) shading indicates the values greater than 0.5 (2.0) mm day−1, respectively. (b) Same as in (a), except for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) The 4-h accumulated rainfall [mm (4 h)−1] averaged over the CNTL runs over the 30-day periods before the onsets of monsoon rainfall in AZNM. The monsoon onset dates were adopted from Higgins et al. (1997). The light (heavy) shading indicates the values greater than 0.5 (2.0) mm day−1, respectively. (b) Same as in (a), except for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) The 4-h accumulated rainfall [mm (4 h)−1] averaged over the CNTL runs over the 30-day periods before the onsets of monsoon rainfall in AZNM. The monsoon onset dates were adopted from Higgins et al. (1997). The light (heavy) shading indicates the values greater than 0.5 (2.0) mm day−1, respectively. (b) Same as in (a), except for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Diurnal variations in the preonset winds (m s−1) at the lowest model layer averaged over the CNTL runs. The light and heavy shading indicates the values greater than 4 and 6 m s−1, respectively.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Diurnal variations in the preonset winds (m s−1) at the lowest model layer averaged over the CNTL runs. The light and heavy shading indicates the values greater than 4 and 6 m s−1, respectively.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Diurnal variations in the preonset winds (m s−1) at the lowest model layer averaged over the CNTL runs. The light and heavy shading indicates the values greater than 4 and 6 m s−1, respectively.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Same as in Fig. 6, but water vapor flux [(g kg−1)(m s−1)] during the preonset period. The light (heavy) shading indicates the magnitude of the lowest-layer water vapor flux exceeding 40 (60) (g kg−1)(m s−1). (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Same as in Fig. 6, but water vapor flux [(g kg−1)(m s−1)] during the preonset period. The light (heavy) shading indicates the magnitude of the lowest-layer water vapor flux exceeding 40 (60) (g kg−1)(m s−1). (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Same as in Fig. 6, but water vapor flux [(g kg−1)(m s−1)] during the preonset period. The light (heavy) shading indicates the magnitude of the lowest-layer water vapor flux exceeding 40 (60) (g kg−1)(m s−1). (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Evaporation differences between the 30-day averages over the postonset and preonset periods: (a) averages over all six years, (b) averages over three wet years, and (c) averages over two dry years. The light (heavy) shading indicates the values smaller (greater) than −0.2 (0.2) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Evaporation differences between the 30-day averages over the postonset and preonset periods: (a) averages over all six years, (b) averages over three wet years, and (c) averages over two dry years. The light (heavy) shading indicates the values smaller (greater) than −0.2 (0.2) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Evaporation differences between the 30-day averages over the postonset and preonset periods: (a) averages over all six years, (b) averages over three wet years, and (c) averages over two dry years. The light (heavy) shading indicates the values smaller (greater) than −0.2 (0.2) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Similar to Fig. 9, but for lowest model-layer specific humidity (g kg−1). The light (heavy) shading indicates the values greater than 1.0 (2.0) g kg−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Similar to Fig. 9, but for lowest model-layer specific humidity (g kg−1). The light (heavy) shading indicates the values greater than 1.0 (2.0) g kg−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Similar to Fig. 9, but for lowest model-layer specific humidity (g kg−1). The light (heavy) shading indicates the values greater than 1.0 (2.0) g kg−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Differences in diurnal variations in rainfall [mm (4 h)−1] between the GC6 and CNTL (GC6 − CNTL) runs during the preonset period. The light (dark) shading indicates the values smaller (greater) than −0.1 (0.1) mm (4 h)−1. (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Differences in diurnal variations in rainfall [mm (4 h)−1] between the GC6 and CNTL (GC6 − CNTL) runs during the preonset period. The light (dark) shading indicates the values smaller (greater) than −0.1 (0.1) mm (4 h)−1. (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Differences in diurnal variations in rainfall [mm (4 h)−1] between the GC6 and CNTL (GC6 − CNTL) runs during the preonset period. The light (dark) shading indicates the values smaller (greater) than −0.1 (0.1) mm (4 h)−1. (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Differences in the lowest model-layer water vapor fluxes [(g kg−1)(m s−1)] between the GC6 and CNTL (GC6 − CNTL) during the preonset period. The light (dark) shading indicates the values greater 4 (8) (g kg−1)(m s−1). (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Differences in the lowest model-layer water vapor fluxes [(g kg−1)(m s−1)] between the GC6 and CNTL (GC6 − CNTL) during the preonset period. The light (dark) shading indicates the values greater 4 (8) (g kg−1)(m s−1). (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
(a) Differences in the lowest model-layer water vapor fluxes [(g kg−1)(m s−1)] between the GC6 and CNTL (GC6 − CNTL) during the preonset period. The light (dark) shading indicates the values greater 4 (8) (g kg−1)(m s−1). (b) Same as in (a), but for the postonset period.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Differences in the lowest model-layer winds (m s−1) between the GC6 and CNTL (GC6 − CNTL) in the postonset period. The light (dark) shading indicates the values exceeding 0.4 (0.8) m s−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Differences in the lowest model-layer winds (m s−1) between the GC6 and CNTL (GC6 − CNTL) in the postonset period. The light (dark) shading indicates the values exceeding 0.4 (0.8) m s−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Differences in the lowest model-layer winds (m s−1) between the GC6 and CNTL (GC6 − CNTL) in the postonset period. The light (dark) shading indicates the values exceeding 0.4 (0.8) m s−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Differences in evaporation (mm day−1) averaged over the 30-day periods before and after the onset of monsoon rainfall in AZNM between the GC6 and CNTL (GC6 − CNTL). The light (dark) shading indicates the values greater than 1.0 (4.0) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Differences in evaporation (mm day−1) averaged over the 30-day periods before and after the onset of monsoon rainfall in AZNM between the GC6 and CNTL (GC6 − CNTL). The light (dark) shading indicates the values greater than 1.0 (4.0) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Differences in evaporation (mm day−1) averaged over the 30-day periods before and after the onset of monsoon rainfall in AZNM between the GC6 and CNTL (GC6 − CNTL). The light (dark) shading indicates the values greater than 1.0 (4.0) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The 6-yr average monthly rainfall (mm day−1) in the GC6 runs. The light (dark) shading indicates the values greater than 1.0 (2.5) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The 6-yr average monthly rainfall (mm day−1) in the GC6 runs. The light (dark) shading indicates the values greater than 1.0 (2.5) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
The 6-yr average monthly rainfall (mm day−1) in the GC6 runs. The light (dark) shading indicates the values greater than 1.0 (2.5) mm day−1.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Daily rainfall variations in the AZNM region from the GC6 (solid line) and the CNTL (dashed line). The vertical lines indicate the monsoon rainfall onset dates determined by Higgins et al. (1997) from rain gauge analysis data.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Daily rainfall variations in the AZNM region from the GC6 (solid line) and the CNTL (dashed line). The vertical lines indicate the monsoon rainfall onset dates determined by Higgins et al. (1997) from rain gauge analysis data.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
Daily rainfall variations in the AZNM region from the GC6 (solid line) and the CNTL (dashed line). The vertical lines indicate the monsoon rainfall onset dates determined by Higgins et al. (1997) from rain gauge analysis data.
Citation: Journal of Climate 18, 23; 10.1175/JCLI3584.1
