Current Research Activities

Matt Newman (

o Examining the sensitivity of low frequency variability to the time of year
o Extreme springtime weather events over North America and their relation to remote sources of forcing
o Applying stochastic forcing in models of low frequency variability

  1. The sensitivity of low frequency variability to the time of year

    Over the Pacific basin, the annual cycle of the 300 mb zonal winds from January through June has three main characteristics:

    o The North Pacific jet, which is strongest in late January, weakens by about half by
    June. As it does so, the jet also moves poleward (particularly east of the
    dateline, where it rapidly moves poleward about 15o in latitude during
    mid-March), extends eastward an additional 10o in longitude, and narrows its
    meridional width.
    o The subtropical jet forms abruptly in mid-March (at about the same time the Pacific
    jet moves poleward). It extends southwestward from the southern U.S. to a point
    near the dateline at about 15oN, peaks in April and May but remains throughout
    the summer.
    o The zero wind line, which holds relatively steady from December through April, moves
    rapidly northward by about 15o latitude during late May and June, with
    the tropical belt of easterlies extending to all longitudes by mid-June.

    The associated monthly mean absolute vorticity gradient can be seen in Figure 1.

    Figure 1. Monthly mean absolute vorticity gradient at 300 mb for the period 1980-1993 (from NMC
    observations). Contour interval is 7.5 x 10-5 s-1. Values greater than 30 x 10-5 s-1 are shaded
    in blue. Negative values are shaded in red and indicated with dashed contours. The thick black line
    represents the zero wind line.

    The two center figures are Hovmuller plots (time vs. latitude) of the annual cycle of absolute vorticity
    gradient at 300 mb, zonally averaged between 135o E-180o (top) and 180o-135o W (bottom). Contour interval
    is 7.5 x 10-5 s-1. Negative values are red; darker blue shading indicates larger values.

    Given such relatively rapid changes in mean flow structure and intensity at certain times of
    the year, one might reasonably expect Rossby wave propagation and dispersion characteristics
    also to be affected, particularly over the North Pacific/North America area within and
    downstream of the Pacific waveguide. For example, even though the zonal wind speed over the
    east Pacific has decreased by half during spring, the vorticity gradient is increased.
    From simple Rossby wave theory, then, we expect that the horizontal scale of disturbances
    should be reduced.

    This is indeed the case, which can be shown by computing forcing influence functions,
    which are essentially the gradient of the regional forced response with respect to the forcing
    pattern. In other words, an influence function is a map showing regions where forcing is
    particularly effective in producing a response in a geographical target area. For our
    barotropic model, such a map will depend upon the base state winds, since the propagation
    of forced Rossby waves will depend upon the local ratio of the zonal wind speed and absolute
    vorticity gradient.

    For example, Figure 2 shows the influence function for each month, when the target area chosen
    is 30o in diameter and is centered over the western U.S. (40o N latitude, 115o W longitude).
    Also included in the figure is a histogram of the pattern correlation between the influence
    functions of consecutive months. Thus, for example, J-F indicates the pattern correlation
    between the January influence function and the February influence function. A value of 1 in
    this plot means that the influence function is unchanged (apart possibly from a constant
    multiplicative factor) between the two months.

    In general, as time passes from January to June the sensitivity of western U.S. height
    anomalies to forcing in the west Pacific increases and the sensitivity to forcing in the
    tropical east Pacific decreases. Distinct month-to-month change is particularly noticeable
    from March through July. The geographical scale of the pattern over the west Pacific steadily
    decreases from one month to the next, resulting in lower pattern correlation between spring
    months. Although the overall sensitivity declines from March to May, as one might perhaps
    expect from the weakening of the Pacific jet, it increases in June, when the extrema have
    larger amplitude than at any time in winter. (Actually, examination of the influence functions
    computed using semi-monthly means shows that the period of greatest sensitivity is from mid-May
    to mid-June.)

    Since the total change in the response is due to the projection of the actual forcing on the
    influence function, it is clear that the geographical scale of the influence function can be as
    important as its amplitude. For example, consider a region of forcing south of the
    30o N latitude circle, all of one sign, between 120o E and 180o. In March,
    this will produce a strong response over the western U.S., since the influence function in this
    region is all of one sign. However, in June, this region of forcing will be too large:
    forcing at 135o E will be offset by forcing at 160o E. Thus, although the
    western U.S. response to June steady forcing can be large, the details of the forcing pattern
    may be more important than during the winter months. Even if the region of forcing is not too
    large, the reduced month-to-month pattern consistency of the influence function in the spring
    months implies that applying the same steady forcing throughout spring could nevertheless
    produce a non-steady response over North America. For example, Figure 2 suggests that steady forcing
    over Southeast Asia will produce an increase in western U.S. height in April, a decrease in
    May, and an increase again in June. A similar effect would occur during fall.

    Figure 2. The influence function for the target area centered at 40o N latitude, 115o W longitude. The center
    of the target area is indicated by a large filled circle. The shading interval is 6 x 1011 ms-2; values between
    -6 x 1011 ms-2 and 6 x 1011 ms-2 are unshaded.

    Newman, M. and P. D. Sardeshmukh, 1996: The impact of the annual cycle upon the North Pacific/North American response to low frequency forcing. Submitted to J. Atmos. Sci.

  2. Extreme springtime weather events over North America

    Traditionally in climate dynamics, much attention has been focused on understanding low
    frequency variability during the Northern Hemisphere winter. This is perhaps in part because
    this is the time of peak synoptic scale activity. However, low frequency variability has
    substantial amplitude throughout the year. Furthermore, the societal impact of extreme weather
    events is not limited to the winter. Heat waves can have a substantial affect upon both health
    and energy consumption, and the agricultural impact of drought is well known.

    As can be seen in the previous section, height anomalies over the western U.S. may have
    enhanced sensitivity to west Pacific forcing during May and June. To what extent is this
    observed? To help answer this question, Figure 3 shows some results of an
    analysis of observed divergence and streamfunction fields at 200 mb, for the period May 15-June
    15 of the years 1980-1993. On the left is global divergence linearly regressed against
    streamfunction over south-central Canada, for lags of -12, -6, and 0 days prior to maximum
    streamfunction amplitude. On the right is global streamfunction regressed against divergence
    over the Philippines, for lags of -6, 0, and +6 days prior to maximum divergence amplitude.
    (Similar results can be obtained if OLR is used instead of divergence.) These results indicate
    that the theoretical region of sensitivity to forcing is in fact important in the late spring.
    At the time of year when the southeast Asian monsoon is intensifying, anomalies in west Pacific
    convection are followed about a week later by large streamfunction anomalies over North America.

    Figure 3. Linear regression results, at various lags. Left: global 200 mb divergence patterns from
    regression against 200 mb streamfunction over south-central Canada (averaged between 52.6oN-58.1oN
    latitude and 95.6oW-101.2oW longitude). Right: global 200 mb streamfunction patterns from regression
    against 200 mb divergence over the Philippines (averaged between 8.3oN-13.8oN latitude
    and 118.1oE-123.8oE longitude).
    Positive contours are blue; negative are red. Contour interval for the divergence plots (left) is
    .3 X 10-6 s-1; for the streamfunction plots (right) 106 m2/s. Green shading indicates the 99% confidence
    level (correlation magnitude greater than .3).
    o1988 U.S. drought

    Of particular interest is the drought that occurred over much of the contiguous United States
    during the spring and summer of 1988. Overall damages to the society and environment made the
    1988 drought one of the nation's greatest disasters of the twentieth century. In April-June
    averages of upper-level geopotential height, the drought was associated with what appears to be
    a wavetrain that emanates from the tropics east of the dateline and arcs across North America
    Figure 4). For this reason, it has been suggested that the drought was caused by a northward
    shift in the ITCZ over the east Pacific. This in turn may have been caused by the onset of a La Nina.

    However, this shift in the ITCZ didn't occur until June, whereas the drought started in early
    April. Furthermore, an analysis of the height data for this period shows that the wavetrain is
    mostly an artifact of the averaging period used. (This can be seen in Figure 4, which shows the
    monthly and seasonal mean 200 mb height anomalies for this period.)

    Figure 4. April-May-June mean and April, May, and June monthly mean 200 mb geopotential height anomalies.

    In fact, the April-May-June mean field is quite similar to one obtained by merely dividing the
    June mean by three.

    Further analysis of 10-day lowpass data shows that the June mean was in turn dominated by two
    episodes of rapidly intensifying anticyclones. The development of both these anticyclones
    appears to follow the regression results, with an increase in anomalous convection (and
    associated upper-level divergence) occurring about a week prior to the height maximum over
    north-central North America.

    To demonstrate the extent to which the observed forcing over the west Pacific contributed to
    the growth of the anticyclones, we show the
    results of a barotropic model run (Figure 5),
    also started about a week prior to the onset of the early June anticyclone and using as forcing
    the observed 200 mb divergence over only the southeast Asia/west Pacific region. This simple
    model can nevertheless reproduce the timing and amplitude of the anomaly quite well, although
    the position is in error by 30o to the southwest.

    Figure 5. (Left) Observed 10-day Lanczos filtered streamfunction anomalies at 200 mb. (Right) Modelled
    streamfunction anomalies at 200 mb. Contour interval is 4 x 106 m2/s.

    Thus, it appears that the April-May-June mean height anomaly was primarily due to two
    wavetrains which were forced by anomalous convection over the west Pacific (perhaps related to
    the southeast Asian monsoon) during June. In this scenario, the shift of the ITCZ was perhaps
    not important, although anomalous SST could still play a role. It should again be stressed
    that, although the June anticyclones certainly produced the June heat wave and therefore played
    an important role in enhancing the severity of the drought, they did not initiate the drought.
    Also, the drought extended well into August, at which time height anomalies were again weak,
    suggesting that the role of surface feedbacks could also become important. The causes for the
    very dry spring that preceeded the June heat waves remain to be determined.

    Chen, P. and M. Newman, 1996: Rossby-wave propagation and the rapid development of upper-level anomalous anticyclones during the 1988 U.S. drought. Submitted to J. Climate.

    Chen, P. and M. Newman, 1995: Rapid development of upper-level anticyclones during the 1988 U.S. drought. Proceedings of the Twentieth Annual Climate Diagnostics Workshop, Seattle, Washington.

  3. Stochastic forcing in models of low frequency variability

    Newman, M., P. D. Sardeshmukh, and C. Penland, 1996: Stochastic forcing of the wintertime extratropical flow. J. Atmos. Sci., in press.

    Matt Newman (
    Back to curriculum vitae