1. Introduction
During the Northern Hemisphere (NH) summer, the dominant rainfall regime over the United States is characterized by an inverse relationship between rainfall anomalies over the Southwest including Arizona and New Mexico (AZNM) and anomalies over the Great Plains (Mo et al. 1997; Higgins et al. 1997). This rainfall dipole pattern is partially explained by the low-level moisture transport. The low-level jet (LLJ) (Wang and Paegle 1996; Helfand and Schubert 1995) over the Great Plains transports moisture from the Gulf of Mexico to the central United States and moisture flux convergence downwind from the jet core enhances rainfall. When the LLJ transports more moisture to the central United States, less moisture is transported to AZNM. That implies wetness over the Great Plains and dryness over AZNM. In addition to the moisture supply from the Gulf of Mexico, rainfall over AZNM is also supported by moisture sources from the Gulf of California (Adams and Comrie 1997). Many studies examined the influence of boundary forcing and soil moisture on summer rainfall regimes and the role played by the LLJ on interannual timescales (Ting and Wang 1997; Lau and Peng 1992; Trenberth and Branstator 1992). However, there is little emphasis on the intraseasonal rainfall variability.
During winter, the rainfall regimes over the Southwest are modulated by two modes in the intraseasonal band (Mo 1999). One mode has a spectral peak near 40 days and is associated with the Madden–Julian Oscillation (MJO) (Madden and Julian 1972, 1994). The other oscillatory mode has a period of 22–28 days (22-day mode). The composites of outgoing longwave radiation anomalies (OLRAs) associated with this 22-day mode reveal that cloud bands propagate northward from the eastern Pacific just north of the ITCZ, through California to the Pacific Northwest. In the Tropics, OLRAs propagate westward from the central Pacific through the western Pacific to the Indian Ocean.
During the NH summer, the MJO is known to modulate the Indian monsoon onset and the Asian Mei-yu development (Lau and Chan 1986). In addition to the MJO, a 10–20-day mode was found in the westward propagating waves associated with the Indian monsoon. Together with the MJO, they determine the monsoon ridge positions and regulate the monsoon wet and dry periods (Krishnamurti et al. 1985).
The present paper examines the impact of intraseasonal oscillations on the summer precipitation regimes over North America. Singular spectrum analysis (SSA;Vautard and Ghil 1989; Vautard et al. 1992) is used to determine the oscillatory modes. The same method was used by Mo (1999) to study the winter precipitation regimes in the Southwest. Evidence will show that the intraseasonal mode with a period of 22–25 days (22-day mode) regulates the summer precipitation regime. The life cycle of the summer 22-day mode is documented and its linkages to the Tropics are discussed. The datasets and procedures used are described in section 2. Evidence that intraseasonal oscillations modulate precipitation over AZNM and the Great Plains is presented in section 3. The 22-day mode is examined in section 4. Conclusions are given in section 5.
2. Data and procedures
The data used in this study were global gridded analyses from the National Centers for Environmental Prediction–National Center for Atmospheric Research 40-yr reanalysis (Kalnay et al. 1996). The data are on a 2.5° lat × 2.5° long grid and cover the period from 1 January 1968 to 31 December 1997. Daily averages of the National Oceanic and Atmospheric (NOAA) satellite outgoing longwave radiation (OLR) data (Liebmann and Smith 1996) were used as a proxy for tropical convection. The OLR data cover the period from 1 January 1979 to 31 December 1997.
The seasonal cycle at each grid point is defined as the grand mean plus the first and second harmonics with periods of 12 and 6 months, respectively. The difference between the field and the seasonal cycle is defined as the anomaly at that grid point. To obtain the intraseasonal signal, data were filtered using the minimum bias window developed by Papoulis (1973) to retain periods in the range of 10–90 days.
Over the United States, daily observed precipitation derived from gridded hourly station data (Higgins et al. 1996) was used to obtain precipitation composites. The data are on a 2.0° lat × 2.5° long grid covering the period from 1963 to 1995. Daily mean climatology was obtained for the entire period and was smoothed using a 7-day running average. Daily precipitation anomalies are defined as departures from the smoothed mean daily climatology.
SSA was used to determine oscillatory modes in time series. SSA is basically a statistical technique related to empirical orthogonal function (EOF) analysis, but in the time space (Vautard and Ghil 1989; Vautard et al. 1992). Quasi-periodic signals appear as pairs of degenerate eigenmodes and their corresponding eigenfunctions in the time domain (T-EOFs) are in quadrature with each other. This means that the maximum correlation between the two T-EOFs is higher than 0.9 and the lag at which the maximum occurs is about ¼ of the period. The original time series can be projected onto T-EOFs to obtain principal components in the time domain (T-PCs). The same procedure was used to study the intraseasonal modulation of the winter precipitation regimes over the western region of the United States (Mo 1999). For details, readers are referred to Vautard and Ghil (1989). A window length of 61 days was used to highlight oscillations on intraseasonal timescales. Results are not sensitive to the particular window length used. The SSA modes are not pure sines and cosines. The dominant periods of T-PCs were estimated using a Blackman–Tukey analysis with a bandwidth of 0.0074.
3. Intraseasonal modulation of precipitation over the United States
Rainfall does not obey a normal distribution and data are not long enough to determine a rainfall distribution function needed to compute the covariance matrix for SSA. OLRAs may not be a good representation of rainfall in midlatitudes. Therefore, the 200-hPa divergence field is used in this study. An example is given in Fig. 1. It shows the 200-hPa divergence and rainfall anomalies averaged from 32° to 36°N for June–September 1979. Both fields were smoothed by a 7-day running average. Positive (negative) rainfall anomalies correspond to anomalous divergence (convergence). In June and July 1979, both fields show stationary and eastward-moving components. The eastward propagation of anomalies was stronger and more regular in August and September. Negative (positive) anomalies over the central United States (80°–100°W) are often accompanied with positive (negative) anomalies over the Southwest (100°–110°W). The average interval between wet periods (anomalous divergence) over the Southwest is roughly about 28 days
To examine intraseasonal variability, the AZNM D200 index was constructed by averaging the 10–90-day-filtered 200-hPa divergence over AZNM (107.5°–112.5°W, 32°–36°N). The rainfall and OLRA indices were constructed the same way. The rainfall index was smoothed by a 7-day running average, but it was not 10–90-day filtered. To study rainfall events over the Great Plains, the Great Plains indices for OLRAs, 200- hPa divergence, and rainfall anomalies were formed by averaging the anomalies over the area (85°–105°W, 32°–40°N).
Positive and negative events were selected according to the threshold criterion. The standard deviation of the AZNM D200 index for June–September (JJAS) was computed. A positive event starts when the index is above 1.2 standard deviations. That date is defined as the onset date. The event ends when the index drops below the threshold. Negative events can be selected the same way.
Lagged composites of the 10–90-day-filtered 200-hPa divergence and rainfall anomalies over the United States were computed for positive and negative events for JJAS from 20 days before to 20 days after onset. The evolution of the 200-hPa divergence difference between positive and negative events (Fig. 2) shows that a three-cell pattern propagates eastward from the North Pacific through AZNM, the Great Plains, to the eastern United States. Anomalies over AZNM are negative at day −10. They move eastward slowly from day −10 to day −6, but the speed increases after day −6. Positive anomalies are located over AZNM at day −2 and reach a maximum at day 2. After day 2, negative anomalies move from the west coast of California to AZNM, while positive anomalies move toward the eastern United States. At day 10, negative anomalies are located over AZNM. The wet and dry cycle of the AZNM monsoon rainfall estimated from the 200-hPa divergence composites is on average about 20–28 days. These findings can be summed up in a time-longitude cross section (Hovmöller diagram) of the 200-hPa divergence difference between positive and negative events averaged from 32° to 36°N (Fig. 3b). It has both stationary and eastward moving components. Anomalies propagate from the Southwest to the eastern United States and there is a phase reversal between anomalies over AZNM and the Great Plains.
The above results are confirmed by the rainfall composites keyed to the same AZNM D200 index (Fig. 3a). There is a good correspondence between positive (negative) rainfall anomalies and the 200-hPa anomalous divergence (convergence). The resemblance between the rainfall and 200-hPa divergence differences suggests that the 200-hPa divergence is a good representation of the low-frequency rainfall signal.
The above results are reproduced in a longer rainfall dataset (Higgins et al. 1996) covering the period from 1963 to 1995. For each season, the seasonal mean anomaly from June to September was removed from rainfall daily anomalies to concentrate on the intraseasonal band. The AZNM rainfall index was formed by averaging rainfall anomalies over the same AZNM area. A positive (negative) event starts when the AZNM rainfall index is above 85th (below 15th) percentile. Composites of rainfall and the 10–90-day filtered 200-hPa divergence were computed from 20 days before to 20 days after the onset for summer. The time–longitude cross-section plots of difference fields (3c and 3d) averaged from 32° to 36°N should be compared with Figs. 3a and 3b. The magnitudes of anomalies are weaker because the rainfall data are noisier. There are differences among composites, but key features such as the eastward propagation of rainfall anomalies from AZNM to the Great Plains and an inverse relationship between anomalies over the two regions are well reproduced. In each case, rainfall anomalies over AZNM change sign twice within 30 days. This suggests that rainfall over AZNM and the Great Plains is modulated by intraseasonal oscillations.
4. Oscillatory modes
a. SSA results
In this section, SSA was performed on the time series of the AZNM D200 index to determine the periods of oscillations. SSA determines EOFs in the time space (T-EOFs). Then, the time series of the AZNM D200 index was projected on the selected T-EOFs to get the corresponding T-PCs.
The leading mode has a period of 22–26 days, which explains 30% of the variance of the AZNM D200 index. The corresponding T-EOFs are given in Fig. 4a. This mode is referred to as the 22-day mode. The second mode explains about 22% of the variance of the AZNM D200 index. The T-EOFs (Fig. 4b) are not pure sines and cosines. They have a 40-day component, but the corresponding T-PCs also show a weak but statistically significant second peak at about 20 days. The absence of the SSA mode with a period of 40–48 days indicates that the MJO signal is weak. The next mode has a period of 17 days (Fig. 4c). The 17-day mode explains about 16% of the variance. Three modes together explain about 68% of the variance. SSA was also performed on the AZNM OLRA index. The first six T-EOFs are reproduced. The first, second, and third pairs explain about 26%, 21%, and 18% of the variance of the AZNM OLR index, respectively. SSA results are consistent with composites (Fig. 2). Both indicate that the dominant mode is the 22-day mode. The MJO does not play an important role in regulating the AZNM monsoon.
The time series corresponding to these leading modes can be reconstructed based on T-PCs and their corresponding T-EOFs. The reconstructed time series for the 22-day mode (crosses) and the time series of the AZNM D200 index (open circles) are displayed for selected summers in Fig. 5. The summation of the first three modes is also given (dark line). Figure 5 should be compared with the daily AZNM rainfall index (Fig. 6). The rainfall time series plotted are the 7-day running means with the seasonal mean removed, but they are not 10–90-day filtered. These summers are selected because of their strong intraseasonal oscillations.
Notice that the magnitudes of rainfall minima are not as large as the magnitudes of rainfall maxima (Fig. 6). This again indicates that rainfall does not follow a normal distribution. Because of different distribution functions, and smoothing used, the rainfall maxima and minima do not always coincide with the maxima and minima of the 200-hPa divergence index. However, there is a good correspondence between the AZNM D200 index and rainfall index overall. Anomalous divergence (convergence) corresponds with a wet (dry) period. For most years, the 22-day mode dominates the time series. The first three modes capture the phase of the total index well. The 22-day mode explains only about 30% of the variance of the AZNM D200 index, which is not large. However, for some years, it plays a role in regulating the wet and dry periods of the AZNM monsoon.
SSA was performed on the D200 index for the Great Plains. The T-EOFs are the same as those for the AZNM index (Figs. 4a–c). The leading mode is the 22-day mode, which explains about 26% of the variance of the index in the 10–90-day band. This is expected because of the inverse relationship of rainfall anomalies between the two regions. SSA was also performed on the index averaged over the Gulf of Mexico (85°–100°W, 25°–30°N). The leading mode is the 22-day mode. However, the index over Mexico (97.5°–105°W, 15°–25°N) shows only one oscillatory mode and it has a period of about 36–40 days. The composites of the 30–60-day anomalies obtained by Knutson and Weickmann (1987) for the extended summer (May–October) indicate that OLRAs in the area of Central America and Mexico are in phase with OLRAs in the western Pacific, but out of phase with OLRAs in the central Pacific. Enhanced convection in the central Pacific produces enhanced Walker circulation with the descending branch located over Central America and Mexico. This may explain the strong presence of the MJO over Mexico.
b. Composites
Composites of daily rainfall anomalies over the United States, the 10–90-day filtered 200-hPa divergence, OLRAs, 200-hPa streamfunction anomalies, and the vertically integrated moisture fluxes were produced for JJAS based on the reconstructed time series of the AZNM D200 index for the 22-day modes. The threshold is again 1.2 standard deviations of the reconstructed time series. Composites were obtained from 20 days before to 20 days after onset. The statistical significance was assessed by assuming that anomalies obey a normal distribution. There are about 330–350 maps in each composite. Areas where values are statistically significant at the 95% level are shaded. Because rainfall maps are only used for verification, no statistical significance is given.
Figure 7 shows the evolution of the 200-hPa divergence difference keyed to the 22-day mode, which should be compared to Fig. 2. The composite differences keyed to the 22-day mode bear strong resemblance to the differences keyed to the total AZNM D200 index. The magnitudes of anomalies are comparable. This confirms the dominance of the 22-day mode. Both show the eastward propagation of anomalies and a phase reversal between anomalies over AZNM and the Great Plains. Figure 7 shows more regular oscillations and less stationary components. These differences are due to the presence of other modes in the total AZNM index. They contribute to the composites keyed to the total AZNM index (Fig. 2).
The time–longitude cross-section plots for rainfall anomalies and 200-hPa divergence differences are given in Figs. 8a and 8b, respectively. There is a good correspondence between positive (negative) rainfall anomalies and the anomalous 200-hPa divergence (convergence). In comparison to the corresponding plots keyed to the total AZNM D200 index (Figs. 3a and 3b), they show more regular oscillations with a period of 22–25 days. The magnitudes of anomalies are comparable. They all show the eastward propagation of anomalies, and a phase reversal between anomalies over AZNM and the Great Plains.
c. Low-level jet
One important feature associated with precipitation over the central United States and AZNM is the LLJ over the Great Plains (Bonner 1968). The LLJ can be represented by the vertically integrated meridional moisture transport (qv). SSA analysis performed on the (qv) averaged over the Great Plains indicates that the leading mode is the 22-day mode.
Figure 8d shows the time–longitude cross section of the vertical integrated meridional moisture flux (qv) difference averaging from 32° to 36°N between positive and negative events keyed to the 22-day mode. The comparison with Fig. 8a indicates a very good correspondence between the (qv) and rainfall anomalies. The negative (qv) anomalies (95°–85°W), representing the weakening of the LLJ coincide with less rain (negative rainfall anomalies) over the Great Plains, while positive (qv) anomalies coincide with more rain (positive rainfall anomalies). Similar to the rainfall dipole, there is a dipole of (qv) anomalies with centers of action located over the Gulf of Mexico and over the Gulf of California. The similar dipole also appears in the convergence of the moisture flux D(Q) (Fig. 8c).
Figure 9 displays the rainfall difference between positive and negative events and the corresponding D(Q) and moisture flux (qu; qv) at two opposite phases of the 22-day mode. Overall, rainfall anomalies are consistent with the composites of the D(Q) differences. When AZNM is dry (negative rainfall anomalies), there is less moisture transported into AZNM from the Gulf of Mexico and from the Gulf of California (Figs. 9d and 9f). When AZNM is wet (Fig. 9b), increased southerlies are located between a cyclonic–anticyclonic pair (Fig. 9e). The LLJ brings more moisture to AZNM. At the same time, the jet that transports moisture from the Gulf of California to the AZNM area also strengthens.
d. Tropical linkages
Similar to the winter 22-day mode, the summer 22-day mode also has a tropical connection. The OLRA and corresponding 200-hPa streamfunction differences keyed to the summer 22-day mode from day −2 to day 6 are given in Fig. 10. From days −2 to 6, anomalous divergence (positive rain anomalies) moves from AZNM to the central United States (Figs. 7c–e). In the tropical Indian–Pacific sector, OLRA composites show a three-cell pattern propagating westward. Enhanced convection (negative OLRAs) centered near 150°E at day −2, moves to 130°E at day 6, while positive OLRAs located in the western Pacific at day −2 move to the Indian Ocean at day 6. The 200-hPa streamfunction composites also show westward-propagating waves. The wave trains extend from the area of enhanced convection through the North Pacific, the Gulf of Alaska, to North America (Figs. 10d and 10e).
At day −2, OLRAs along the west coast show another three-cell pattern with positive anomalies over the Pacific Northwest and central America and negative anomalies over the Southwest (Fig. 10a). The corresponding 200-hPa streamfunction difference (Fig. 10d) over North America is consistent with OLRAs. At day 2, rainfall over AZNM reaches a maximum (Fig. 9b). In addition to the three-cell OLRA pattern in the Indian–Pacific sector, negative OLRAs are found in the eastern Pacific with positive OLRAs located over Central America. The corresponding 200-hPa streamfunction difference shows negative anomalies over the Southwest and positive anomalies located over the Pacific Northwest and the central United States (Fig. 10e). This is consistent with the composite difference of moisture fluxes. Figure 9e shows increasing southerlies located between the cyclonic–anticyclonic dipole. At day 6, rainfall (anomalous 200-hPa divergence) moves from AZNM into the central United States. The 200-hPa streamfunction anomaly pattern (Fig. 10f) is similar to the composite keyed to wet events over the central United States (Mo et al. 1997).
The westward propagation of the OLRAs and the corresponding westward moving 200-hPa streamfunction anomalies over the Pacific–North American sector are also features of the winter 22-day mode. These anomalies are statistically significant at the 95% level. However, the magnitudes of anomalies are weaker in comparison with the composites keyed to the winter 22-day mode. The westward propagation is more apparent in the time–longitude cross section centered just north of the equator (10°–20°N). OLRAs propagate westward from the central Pacific through the western Pacific to the Indian Ocean (Fig. 11) and the propagation completes a cycle in about 22–25 days. During winter, OLRAs propagate northward from the eastern Pacific through California to the Pacific Northwest. For summer composites, there is a three-cell pattern over the west coast of North America but there is little evidence of northward propagation.
5. Conclusions
Over the United States, the summer rainfall regime is dominated by a dipole pattern with centers of action located over AZNM and the Great Plains. The evolution of the AZNM monsoon based on composites indicates that rainfall anomalies propagate eastward from the North Pacific through AZNM, the Great Plains, to the eastern United States. Anomalies over AZNM change signs twice within 30 days indicating an intraseasonal modulation of rainfall.
The intraseasonal variability is examined further using singular spectrum analysis and composites of OLRAs, 200-hPa divergence and a gridded rainfall dataset. SSA performed on the AZNM index shows that the dominant mode is the 22-day mode. The second mode has an MJO component, but it also has a second spectral peak at 20 days. The third mode is the 17-day mode. Three modes together explain about 68% of the variance in the intraseasonal band. The wet and dry periods of the AZNM monsoon are modulated by the 22-day mode. This is also the dominant mode for rainfall variability over the Great Plains. The influence of the MJO is secondary.
The 22–25-day mode is found in the vertically integrated moisture fluxes over the Great Plains. When AZNM is wet, more moisture is transported from both the Gulf of Mexico and the Gulf of California to AZNM. The situation reverses when the oscillation reaches the other phase. The largest impact of the MJO is on precipitation over Mexico. SSA performed on the 200-hPa divergence and OLRAs averaged over Mexico shows only one oscillatory mode of about 36–40 days.
The 22-day mode is linked to tropical convection. When rainfall associated with the 22-day mode travels eastward from AZNM to the Great Plains, the OLRAs composites show westward propagating waves just north of the equator. When enhanced convection reaches the western Pacific, rainfall diminishes in AZNM. When convection in the western Pacific is suppressed and enhanced convection is located in the central Pacific, rainfall intensifies in AZNM.
During the summer season, intraseasonal oscillations in the Tropics are as strong as their counterparts in winter. The 22-day mode, the MJO and the 17-day mode are also leading oscillatory modes in the Tropics (Ghil and Mo 1991). For both seasons, the MJO is the dominant mode in the Tropics. The second mode is the 22-day mode. In comparison to the winter 22-day mode, the magnitudes of OLRAs in the Tropics associated with the summer 22-day mode are weaker. The ratio of the 10–30-day OLR variance to the 10–90-day OLR variance is only about 35%–50% in the western and the central Pacific for summer. The ratio increases to 50%–60% for the areas in the eastern Pacific, and the Gulf of Mexico and North America. This may explain the weaker magnitudes of OLR anomalies in the Tropics (Fig. 10).
The summer 22-day mode has some features similar to the winter mode. Both 22-day modes show that OLRAs propagate westward in the Tropics just above the equator (10°–20°N). Both modes have impact on precipitation anomalies over the United States. During winter, the impact is felt over the western region of the United States and during summer it modulates the AZNM monsoon and rainfall activities over the Great Plains.
REFERENCES
Adams, D. K., and A. C. Comrie, 1997: The North American monsoon. Bull. Amer. Meteor. Soc.,78, 2197–2231.
Bonner, W. D., 1968: Climatology of the low level jet. Mon. Wea. Rev.,96, 833–850.
Ghil, M., and K. C. Mo, 1991: Intraseasonal oscillations in the global atmosphere. Part I: Northern Hemisphere and Tropics. J. Atmos. Sci.,48, 752–779.
Helfand, H. M., and S. D. Schubert, 1995: Climatology of the Great Plains low level jet and its contribution to the continental moisture budget of the United States. J. Climate,8, 784–806.
Higgins, R. W., J. E. Janowiak, and Y. P. Yao, 1996: A gridded hourly precipitation data base for the United States (1963–1993). NCEP/Climate Prediction Center Atlas No. 1, NCEP/NWS/NOAA, 47 pp.
——, Y. Yao, J. Janowiak, and K. C. Mo, 1997: Influence of the Great Plains low level jet on summertime precipitation and moisture transport over the central United States. J. Climate,10, 481–507.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc.,77, 437–471.
Knutson, T. R., and K. M. Weickmann, 1987: 30–60 day atmospheric oscillations: Composite life cycle of convection and circulation anomalies. Mon. Wea. Rev.,115, 1407–1435.
Krishnamurthi, T. N., P. K. Jayakumar, J. Sheng, N. Surgi, and A. Kumar, 1985: Divergence circulations on the 30–50 day time scale. J. Atmos. Sci.,42, 364–375.
Lau, K.-M., and P. H. Chan, 1986: Aspects of the 40–50 day oscillation during the northern summer as inferred from outgoing long wave radiation. Mon. Wea. Rev.,114, 1354–1367.
——, and L. Peng, 1992: Dynamics of atmospheric teleconnections during the northern summer. J. Climate,5, 140–158.
Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc.,77, 1275–1277.
Madden, R., and P. Julian, 1972: Description of global scale circulation cells in the Tropics with a 40–50 day period. J. Atmos. Sci,29, 1109–1123.
——, and ——, 1994: Observations of the 40–50 day tropical oscillation—A review. Mon. Wea. Rev.,122, 814–837.
Mo, K. C., 1999: Alternating wet and dry episodes over California and intraseasonal oscillations. Mon. Wea. Rev.,127, 2759–2776.
——, J. N. Paegle, and R. W. Higgins, 1997: Atmospheric processes associated with summer floods and droughts in the central United States. J. Climate,10, 3028–3046.
Papoulis, A., 1973: Minimum bias windows for high resolution spectral estimates. IEEE Trans. Inf. Theory,19, 9–12.
Ting, M. F., and H. Wang, 1997: Summertime U. S. precipitation variability and its relation to Pacific sea surface temperature. J. Climate,10, 1853–1873.
Trenberth, K. E., and G. Branstator, 1992: Issues in establishing causes of the 1988 drought over North America. J. Climate,5, 159–172.
Vautard, R., and M. Ghil, 1989: Singular spectrum analysis in nonlinear dynamics with applications to paleoclimatic time series. Physica,35 (D), 392–424.
——, P. Yiou, and M. Ghil, 1992: Singular spectrum analysis: A toolkit for short, noisy chaotic signals. Physica,58 (D), 95–126.
Wang, M., and J. Paegle, 1996: Impact of analysis uncertainty upon regional atmospheric moisture flux. J. Geophys. Res.,101, 7291–7303.
(a) Time–longitude of cross section (Hovmöller diagram) of daily rainfall anomalies averaged from 32° to 36°N for Jun–Sept 1979. Anomalies are smoothed by a 7-day running average. Contour interval is 1.6 mm day−1. Zero contours are omitted. Contours −0.8 and 0.8 mm day−1 are added. Positive values are shaded. (b) Same as (a) but for daily 200-hPa anomalous divergence. Contour interval is 2 × 10−6 s−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) Time–longitude of cross section (Hovmöller diagram) of daily rainfall anomalies averaged from 32° to 36°N for Jun–Sept 1979. Anomalies are smoothed by a 7-day running average. Contour interval is 1.6 mm day−1. Zero contours are omitted. Contours −0.8 and 0.8 mm day−1 are added. Positive values are shaded. (b) Same as (a) but for daily 200-hPa anomalous divergence. Contour interval is 2 × 10−6 s−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) Time–longitude of cross section (Hovmöller diagram) of daily rainfall anomalies averaged from 32° to 36°N for Jun–Sept 1979. Anomalies are smoothed by a 7-day running average. Contour interval is 1.6 mm day−1. Zero contours are omitted. Contours −0.8 and 0.8 mm day−1 are added. Positive values are shaded. (b) Same as (a) but for daily 200-hPa anomalous divergence. Contour interval is 2 × 10−6 s−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Map sequence of the 200-hPa divergence composite difference between positive and negative events keyed to the AZNM D200 index for (a) day −10, (b) day −6, (c) day −2, (d) day 2, (e) day 6, and (f) day 10. Contour interval is 1 × 10−6 s−1. Zero contours are omitted. Contours −0.5 × 10−6 s−1 and 0.5 × 10−6 s−1 are added. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light).
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Map sequence of the 200-hPa divergence composite difference between positive and negative events keyed to the AZNM D200 index for (a) day −10, (b) day −6, (c) day −2, (d) day 2, (e) day 6, and (f) day 10. Contour interval is 1 × 10−6 s−1. Zero contours are omitted. Contours −0.5 × 10−6 s−1 and 0.5 × 10−6 s−1 are added. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light).
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Map sequence of the 200-hPa divergence composite difference between positive and negative events keyed to the AZNM D200 index for (a) day −10, (b) day −6, (c) day −2, (d) day 2, (e) day 6, and (f) day 10. Contour interval is 1 × 10−6 s−1. Zero contours are omitted. Contours −0.5 × 10−6 s−1 and 0.5 × 10−6 s−1 are added. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light).
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) Time–longitude of cross section (Hovmöller diagram) of daily rainfall composite difference averaged from 32° to 36°N between positive and negative events from 20 days before to 20 days after onset keyed to (a) the AZNM D200 index. Contour interval is 0.5 mm day−1. Negative anomalies are shaded and zero contours are omitted. Contours −0.3 and 0.3 mm day−1 are added. (b) Same as (a) but for the 200-hPa divergence. Contour interval is 1 × 10−6 s−1. Contours −0.5 × 10−6 s−1 and 0.5 × 10−6 s−1 are added. (c) Same as (a) but keyed to the rainfall index for the period 1963–95. (d) Same as (b) but keyed to the rainfall index.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) Time–longitude of cross section (Hovmöller diagram) of daily rainfall composite difference averaged from 32° to 36°N between positive and negative events from 20 days before to 20 days after onset keyed to (a) the AZNM D200 index. Contour interval is 0.5 mm day−1. Negative anomalies are shaded and zero contours are omitted. Contours −0.3 and 0.3 mm day−1 are added. (b) Same as (a) but for the 200-hPa divergence. Contour interval is 1 × 10−6 s−1. Contours −0.5 × 10−6 s−1 and 0.5 × 10−6 s−1 are added. (c) Same as (a) but keyed to the rainfall index for the period 1963–95. (d) Same as (b) but keyed to the rainfall index.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) Time–longitude of cross section (Hovmöller diagram) of daily rainfall composite difference averaged from 32° to 36°N between positive and negative events from 20 days before to 20 days after onset keyed to (a) the AZNM D200 index. Contour interval is 0.5 mm day−1. Negative anomalies are shaded and zero contours are omitted. Contours −0.3 and 0.3 mm day−1 are added. (b) Same as (a) but for the 200-hPa divergence. Contour interval is 1 × 10−6 s−1. Contours −0.5 × 10−6 s−1 and 0.5 × 10−6 s−1 are added. (c) Same as (a) but keyed to the rainfall index for the period 1963–95. (d) Same as (b) but keyed to the rainfall index.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) T-EOFs 1 and 2, (b) T-EOFs 3 and 4, and (c) T-EOFs 5 and 6 based on SSA analysis on the AZNM D200 index.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) T-EOFs 1 and 2, (b) T-EOFs 3 and 4, and (c) T-EOFs 5 and 6 based on SSA analysis on the AZNM D200 index.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) T-EOFs 1 and 2, (b) T-EOFs 3 and 4, and (c) T-EOFs 5 and 6 based on SSA analysis on the AZNM D200 index.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Reconstructed time series based on the AZNM 22 day mode (crosses), the sum of the first three modes (dark line) and the AZNM D200 index (open circles) for the (a) 1979, (b) 1981, and (c) 1995 summer. The unit is 1 × 10−6 s−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Reconstructed time series based on the AZNM 22 day mode (crosses), the sum of the first three modes (dark line) and the AZNM D200 index (open circles) for the (a) 1979, (b) 1981, and (c) 1995 summer. The unit is 1 × 10−6 s−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Reconstructed time series based on the AZNM 22 day mode (crosses), the sum of the first three modes (dark line) and the AZNM D200 index (open circles) for the (a) 1979, (b) 1981, and (c) 1995 summer. The unit is 1 × 10−6 s−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
The 7-day running mean precipitation averaged over AZNM based on daily precipitation analysis by Higgins et al. (1996) for the (a) 1979, (b) 1981, and (c) 1995 summer. The unit is mm day−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
The 7-day running mean precipitation averaged over AZNM based on daily precipitation analysis by Higgins et al. (1996) for the (a) 1979, (b) 1981, and (c) 1995 summer. The unit is mm day−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
The 7-day running mean precipitation averaged over AZNM based on daily precipitation analysis by Higgins et al. (1996) for the (a) 1979, (b) 1981, and (c) 1995 summer. The unit is mm day−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Same as Fig. 2 but keyed to the 22-day mode.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Same as Fig. 2 but keyed to the 22-day mode.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Same as Fig. 2 but keyed to the 22-day mode.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Time–longitude cross section of the rainfall difference averaged from 32° to 36°N between positive and negative events from 20 days before to 20 days after onset keyed to the 22-day mode. Contour interval is 0.5 mm day−1. Contours 0.3 and −0.3 mm day−1 are added. (b) Same as (b) but for the 200-hPa divergence. Contour interval 1 × 10−6 s−1. (c) Same as (a) but for the moisture flux divergence D(Q). Contour interval 0.5 mm day−1. (d) Same as (a) but for the vertically integrated meridional moisture flux (qv). Contour interval is 10 kg (m s)−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Time–longitude cross section of the rainfall difference averaged from 32° to 36°N between positive and negative events from 20 days before to 20 days after onset keyed to the 22-day mode. Contour interval is 0.5 mm day−1. Contours 0.3 and −0.3 mm day−1 are added. (b) Same as (b) but for the 200-hPa divergence. Contour interval 1 × 10−6 s−1. (c) Same as (a) but for the moisture flux divergence D(Q). Contour interval 0.5 mm day−1. (d) Same as (a) but for the vertically integrated meridional moisture flux (qv). Contour interval is 10 kg (m s)−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Time–longitude cross section of the rainfall difference averaged from 32° to 36°N between positive and negative events from 20 days before to 20 days after onset keyed to the 22-day mode. Contour interval is 0.5 mm day−1. Contours 0.3 and −0.3 mm day−1 are added. (b) Same as (b) but for the 200-hPa divergence. Contour interval 1 × 10−6 s−1. (c) Same as (a) but for the moisture flux divergence D(Q). Contour interval 0.5 mm day−1. (d) Same as (a) but for the vertically integrated meridional moisture flux (qv). Contour interval is 10 kg (m s)−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) Rainfall composite difference between positive and negative events for day −10, keyed to the reconstructed time series based on the 22-day mode. Contour interval is 0.4 mm day−1. Negative values are shaded. (b) Same as (a) but for day 2, and (c) same as (a) but for day 12, (d) vertically integrated moisture flux difference (qu; qv) (arrows) and moisture flux divergence difference (contours) between positive and negative events averaged from days −11 to −9. Contour interval for D(Q) is 0.4 mm day−1. Negative values are shaded. (e) Same as (d) but for the average from days 1 to 3. Contour interval is 0.5 mm day−1. (f) Same as (d) but for the average from days 11 to 13. The unit for the moisture flux is 10 kg (m s)−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) Rainfall composite difference between positive and negative events for day −10, keyed to the reconstructed time series based on the 22-day mode. Contour interval is 0.4 mm day−1. Negative values are shaded. (b) Same as (a) but for day 2, and (c) same as (a) but for day 12, (d) vertically integrated moisture flux difference (qu; qv) (arrows) and moisture flux divergence difference (contours) between positive and negative events averaged from days −11 to −9. Contour interval for D(Q) is 0.4 mm day−1. Negative values are shaded. (e) Same as (d) but for the average from days 1 to 3. Contour interval is 0.5 mm day−1. (f) Same as (d) but for the average from days 11 to 13. The unit for the moisture flux is 10 kg (m s)−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) Rainfall composite difference between positive and negative events for day −10, keyed to the reconstructed time series based on the 22-day mode. Contour interval is 0.4 mm day−1. Negative values are shaded. (b) Same as (a) but for day 2, and (c) same as (a) but for day 12, (d) vertically integrated moisture flux difference (qu; qv) (arrows) and moisture flux divergence difference (contours) between positive and negative events averaged from days −11 to −9. Contour interval for D(Q) is 0.4 mm day−1. Negative values are shaded. (e) Same as (d) but for the average from days 1 to 3. Contour interval is 0.5 mm day−1. (f) Same as (d) but for the average from days 11 to 13. The unit for the moisture flux is 10 kg (m s)−1.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) OLRA composite difference between positive and negative events keyed to reconstructed time series based on the 22-day mode for day −2. Contour interval is 4 W m−2. Zero contours are omitted. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light). (b) Same as (a) but for day 2, (c) same as (a) but for day 6, (d) same as (a) but for 200-hPa streamfunction difference. Contour interval is 1 × 106 m2 s−1. (e) Same as (d) but for day 2, and (f) same as (d) but for day 6.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) OLRA composite difference between positive and negative events keyed to reconstructed time series based on the 22-day mode for day −2. Contour interval is 4 W m−2. Zero contours are omitted. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light). (b) Same as (a) but for day 2, (c) same as (a) but for day 6, (d) same as (a) but for 200-hPa streamfunction difference. Contour interval is 1 × 106 m2 s−1. (e) Same as (d) but for day 2, and (f) same as (d) but for day 6.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
(a) OLRA composite difference between positive and negative events keyed to reconstructed time series based on the 22-day mode for day −2. Contour interval is 4 W m−2. Zero contours are omitted. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light). (b) Same as (a) but for day 2, (c) same as (a) but for day 6, (d) same as (a) but for 200-hPa streamfunction difference. Contour interval is 1 × 106 m2 s−1. (e) Same as (d) but for day 2, and (f) same as (d) but for day 6.
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Time–longitude cross section for OLRAs composite difference between positive and negative events averaged from 10° to 20°N from 20 days before to 20 days after onset keyed to the 22-day mode. Contour interval is 4 W m−2. Zero contours are omitted. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light).
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Time–longitude cross section for OLRAs composite difference between positive and negative events averaged from 10° to 20°N from 20 days before to 20 days after onset keyed to the 22-day mode. Contour interval is 4 W m−2. Zero contours are omitted. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light).
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
Time–longitude cross section for OLRAs composite difference between positive and negative events averaged from 10° to 20°N from 20 days before to 20 days after onset keyed to the 22-day mode. Contour interval is 4 W m−2. Zero contours are omitted. Areas where positive (negative) values are statistically significant at the 95% level are shaded dark (light).
Citation: Monthly Weather Review 128, 5; 10.1175/1520-0493(2000)128<1490:IMOSPO>2.0.CO;2
