• Allaart, M. A. F., H. Kelder, and L. C. Heijboer. 1993. On the relation between ozone and potential vorticity. Geophys. Res. Lett. 20:811814.

    • Search Google Scholar
    • Export Citation
  • Beekmann, M., G. Ancellet, and G. Megie. 1994. Climatology of tropospheric ozone in southern Europe and its relation to potential vorticity. J. Geophys. Res. 99:1284112853.

    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B. 1993. CLASS for class. Bull. Amer. Meteor. Soc. 74:16971702.

  • Danielsen, E. F. 1968. Stratospheric–tropospheric exchange based on radioactivity, ozone and potential vorticity. J. Atmos. Sci. 25:502518.

    • Search Google Scholar
    • Export Citation
  • Davis, C., S. Low-Nam, M. A. Shapiro, X. Zou, and A. J. Krueger. 1999. Direct retrieval of wind from Total Ozone Mapping Spectrometer (TOMS) data: Examples from FASTEX. Quart. J. Roy. Meteor. Soc. 125:33753391.

    • Search Google Scholar
    • Export Citation
  • Dudhia, J. 1993. A nonhydrostatic version of the Penn State–NCAR Mesoscale Model: Validation tests and simulation of an Atlantic cyclone and cold front. Mon. Wea. Rev. 121:14931513.

    • Search Google Scholar
    • Export Citation
  • Fusco, A. C. and M. L. Salby. 1999. Interannual variations of total ozone and their relationship to variations of planetary wave activity. J. Climate 12:16191629.

    • Search Google Scholar
    • Export Citation
  • Grell, G. A., J. Dudhia, and D. R. Stauffer. 1994. A description of the Fifth-Generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 138 pp. [Available from UCAR Communications, P. O. Box 3000, Boulder, CO 80307.].

    • Search Google Scholar
    • Export Citation
  • Heath, D. F., A. J. Krueger, A. J. Roder, and B. D. Henderson. 1975. The solar scatter ultraviolet and total ozone mapping spectrometer (SBUV/TOMS) for Nimbus G. Opt. Eng. 14:323331.

    • Search Google Scholar
    • Export Citation
  • Maglaras, G. J., J. S. Waldstreicher, P. J. Kocin, A. F. Gigi, and R. A. Marine. 1995. Winter weather forecasting throughout the eastern United States. Part I: An overview. Wea. Forecasting 10:520.

    • Search Google Scholar
    • Export Citation
  • McPeters, R. D., P. K. Bhartia, A. J. Krueger, and J. R. Herman. 1998. Earth Probe Total Ozone Mapping Spectrometer (TOMS) data products user's guide. NASA Tech. Publ. 1998-206895, 70 pp.

    • Search Google Scholar
    • Export Citation
  • Parrish, D. F. and J. C. Derber. 1992. The National Meteorological Center's spectral statistical-interpolation analysis system. Mon. Wea. Rev. 120:17471763.

    • Search Google Scholar
    • Export Citation
  • Ravetta, F. and G. Ancellet. 2000. Identification of dynamical processes at the tropopause during the decay of a cutoff low using high-resolution airborne lidar ozone measurements. Mon. Wea. Rev. 128:32523267.

    • Search Google Scholar
    • Export Citation
  • Reed, R. J. 1950. The role of vertical motions in ozone–weather relationships. J. Meteor. 7:263267.

  • Shapiro, M. A. 1980. Turbulent mixing within tropopause folds as a mechanism for the exchange of chemical constituents between the stratosphere and troposphere. J. Atmos. Sci. 37:9941004.

    • Search Google Scholar
    • Export Citation
  • Shapiro, M. A., A. J. Krueger, and P. J. Kennedy. 1982. Nowcasting the position and intensity of jet streams using a satellite-borne total ozone mapping spectrometer. Nowcasting, K. Browning, Ed., Academic Press, 137–145.

    • Search Google Scholar
    • Export Citation
  • Uccellini, L. W., P. J. Kocin, R. A. Petersen, C. H. Wash, and K. F. Brill. 1984. The Presidents' Day cyclone of 18–19 February 1979: Synoptic overview and analysis of the subtropical jet streak influencing the pre-cyclogenetic period. Mon. Wea. Rev. 112:3155.

    • Search Google Scholar
    • Export Citation
  • Xiao, Q., X. Zou, and Y-H. Kuo. 2000. Incorporating the SSM/I-derived precipitable water and rainfall rate into a numerical model: A case study for the ERICA IOP-4 cyclone. Mon. Wea. Rev. 128:87108.

    • Search Google Scholar
    • Export Citation
  • Zhang, F., C. Snyder, and R. Rotunno. 2002. Mesoscale predictability of the “surprise” snowstorm of 24–25 January 2000. Mon. Wea. Rev. 130:16171632.

    • Search Google Scholar
    • Export Citation
  • Zou, X. and Y-H. Kuo. 1996. Rainfall assimilation through an optimal control of initial and boundary conditions in a limited-area mesoscale model. Mon. Wea. Rev. 124:28592882.

    • Search Google Scholar
    • Export Citation
  • Zou, X., F. Vandenberghe, M. Pondeca, and Y-H. Kuo. 1997. Introduction to adjoint techniques and the MM5 adjoint modeling system. NCAR Tech. Note TN-435-STR, 110 pp. [Available from NCAR, P. O. Box 3000, Boulder, CO 80307-3000.].

    • Search Google Scholar
    • Export Citation
  • Zou, X., Y-H. Kuo, and S. Low-Nam. 1998. Medium-range prediction of an extratropical oceanic cyclone: Impact of initial state. Mon. Wea. Rev. 126:27372763.

    • Search Google Scholar
    • Export Citation
  • Zou, X., Q. Xiao, A. E. Lipton, and G. D. Modica. 2001. A numerical study of the effect of GOES sounder cloud-cleared brightness temperatures on the prediction of Hurricane Felix. J. Appl. Meteor. 40:3455.

    • Search Google Scholar
    • Export Citation
  • Zupanski, D. and F. Mesinger. 1995. Four-dimensional assimilation of precipitation data. Mon. Wea. Rev. 123:11121126.

  • View in gallery

    (a) The geopotential height (solid) and temperature (dashed) at 500 hPa and (b) the sea level pressure (solid) and surface temperature (dashed) at 1200 UTC 25 Jan 2000. Contour intervals are 50 m for the geopotential height, (a) 2.5 and (b) 5.0 K for the temperature, and 4 hPa for the sea level pressure

  • View in gallery

    (a) TOMS ozone observation locations and (b) observed TOMS ozone field during 1200–1900 UTC 24 Jan 2000, and SLP at 1200 UTC 24 Jan 2000, and (c) MPV based on the NCEP analysis and sea surface pressure at 1200 UTC 24 Jan 2000. Contour interval is (b) 20 DU, (c) 1 PVU, and (b), (c) 4 hPa for the sea level pressure

  • View in gallery

    Scatterplot of MPV derived from the 90-km analysis vs TOMS total ozone observations over the same domain as Fig. 2 for the period 15–26 Jan 2000

  • View in gallery

    Values of (a) α and (b) β assuming O3 = α(MPV) + β, where O3 represents TOMS ozone and MPV is calculated from the three datasets: 90-km analyses, 30-km analyses, and 30-km forecasts

  • View in gallery

    Time series of the correlation coefficient between TOMS ozone and MPV for the three datasets (90-km analyses, 30-km analyses, and 30-km forecasts)

  • View in gallery

    The sea level pressure (solid), vertically integrated cloud water (ICLW) (dotted), and vertically integrated rainwater (IRNW) (shaded) at 1200 UTC 25 Jan 2000 for (a) CTRL, (b) BOTH, and (c) RAD. Contour intervals for SLP are 4.0 hPa and for ICLW are 0.2 mm. Areas where IRNW is greater than 0.1 and 0.3 mm are shaded lightly and heavily, respectively

  • View in gallery

    The PV at 360 K at 1200 UTC 24 Jan 2000 from (a) BOTH and (b) RAD. (c) The differences of PV at 360 K between BOTH and RAD (BOTH − RAD). Contour intervals are (a), (b) 1 and (c) 0.5 PVU.

  • View in gallery

    (a) The temperature at 850 hPa at 1200 UTC 24 Jan 2000 from CTRL. (b) Differences of temperature between RAD and CTRL. (c) Differences of temperature between BOTH and CTRL. (d) Differences of temperature between BOTH and RAD. Contour intervals are (a) 2.5, (b), (c) 0.2, and (d) 0.1 K.

  • View in gallery

    (a) The PV at 360 K (solid lines) and SLP (dotted lines) from BOTH, (b) cross section of PV (thin lines) and potential temperature (thick lines) along the line AB indicated in (a), and (c) the same as (b) but for the PV differences between BOTH and RAD (thin lines) and potential temperature (thick lines) at 1800 UTC 24 Jan 2000

  • View in gallery

    (a) The wind speeds and wind vectors at 300 hPa at 1200 UTC 25 Jan 2000 from CTRL. (b) Differences of wind speed between RAD and CTRL. (c) Differences of wind speed between BOTH and CTRL. (d) Differences of wind speed between BOTH and RAD. Contour intervals are (a) 10 and (b)–(d) 2 m s−1

  • View in gallery

    Same as Fig. 10, but for the temperature at 850 hPa. Contour intervals are (a) 2.5, (b), (c) 0.5, and (d) 0.2 K.

  • View in gallery

    Wind differences between BOTH and CTRL at 300 hPa for (a) 1800 UTC 24 Jan, (b) 0000 UTC 25 Jan, and (c) 0600 UTC 25 Jan. Contour interval is 2.0 m s−1

  • View in gallery

    Same as Fig. 12, but for the wind at 850 hPa

  • View in gallery

    The temperature differences (thin line) between BOTH and CTRL at 850 hPa and SLP (thick line) from BOTH for (a) 1800 UTC 24 Jan, (b) 0000 UTC 25 Jan, and (c) 0600 UTC 25 Jan. Contour interval is 0.5 K for temperature differences and 4 hPa for the sea level pressure.

  • View in gallery

    (a) Variation of the central SLP of the storm and (b) the cyclone track errors from the 90-km forecast experiments

  • View in gallery

    The accumulated total precipitation from 1200 UTC 24 Jan to 1200 UTC 25 Jan 2000 from (a) observations, (b) CTRL, (c) RAD, and (d) BOTH. Contour intervals for the precipitation are (a) 5.0 and (b)–(d) 10.0 mm. The solid dot indicated by “DC” in (a) indicates the position of Washington, DC. Solid circles indicate the precipitation observation sites that are used in Fig. 17. Solid circles and triangles indicate the 15 stations used in the calculation in Fig. 18

  • View in gallery

    Simulated and observed hourly precipitation amounts from 1200 UTC 24 Jan to 0000 UTC 16 Jan 2000 at (a) Camp Pickett, VA, (b) Norfolk, VA, (c) Carthage Hatteras, NC, and (d) Laurinburg, NC

  • View in gallery

    (a) Rms errors and (b) mean bias of hourly precipitation at 15 stations over NC, VA, and MD for which the observed total precipitation from 1200 UTC 24 Jan to 0000 UTC 26 Jan 2000 exceeded 20 mm

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 1353 1349 1243
PDF Downloads 10 10 7

Incorporating TOMS Ozone Measurements into the Prediction of the Washington, D.C., Winter Storm during 24–25 January 2000

View More View Less
  • a The Florida State University, Tallahassee, Florida
  • | b NCEP/EMC, Camp Springs, Maryland
  • | c NOAA/ETL, Boulder, Colorado
  • | d National Center for Atmospheric Research,* Boulder, Colorado
  • | e University of Maryland, Baltimore County, Baltimore, Maryland
© Get Permissions
Full access

Abstract

In this study, a methodology is proposed for incorporating total column ozone data from the Total Ozone Mapping Spectrometer (TOMS) into the initial conditions of a mesoscale prediction model. Based on the strong correlation between vertical mean potential vorticity (MPV) and TOMS ozone (O3) that was found in middle latitudes at both 30- and 90-km resolutions, using either analyses or 24-h model forecasts, a statistical correlation model between O3 and MPV is employed for assimilating TOMS ozone in a four-dimensional variational data assimilation (4DVAR) procedure. A linear relationship between O3 and MPV is first assumed: O3 = α(MPV) + β. The constants α and β are then found by a regression method. The proposed approach of using this simple linear regression model for ozone assimilation is applied to the prediction of the 24–25 January 2000 East Coast winter storm. Three 4DVAR experiments are carried out assimilating TOMS ozone, radiosonde, or both types of observations. It is found that adjustments in model initial conditions assimilating only TOMS ozone data are confined to the upper levels and produce almost no impact on the prediction of the storm development. However, when TOMS ozone data are used together with radiosonde observations, a more rapid deepening of sea level pressure of the simulated storm is observed than with only radiosonde observations. The predicted track of the winter storm is also altered, moving closer to the coast. Using NCEP multisensor hourly rainfall data as verification, the 36-h forecasts with both TOMS ozone and radiosonde observations outperform the radiosonde-only experiments. These results indicate that TOMS ozone data contain valuable meteorological information, which can be used to improve cyclone prediction.

Abstract

In this study, a methodology is proposed for incorporating total column ozone data from the Total Ozone Mapping Spectrometer (TOMS) into the initial conditions of a mesoscale prediction model. Based on the strong correlation between vertical mean potential vorticity (MPV) and TOMS ozone (O3) that was found in middle latitudes at both 30- and 90-km resolutions, using either analyses or 24-h model forecasts, a statistical correlation model between O3 and MPV is employed for assimilating TOMS ozone in a four-dimensional variational data assimilation (4DVAR) procedure. A linear relationship between O3 and MPV is first assumed: O3 = α(MPV) + β. The constants α and β are then found by a regression method. The proposed approach of using this simple linear regression model for ozone assimilation is applied to the prediction of the 24–25 January 2000 East Coast winter storm. Three 4DVAR experiments are carried out assimilating TOMS ozone, radiosonde, or both types of observations. It is found that adjustments in model initial conditions assimilating only TOMS ozone data are confined to the upper levels and produce almost no impact on the prediction of the storm development. However, when TOMS ozone data are used together with radiosonde observations, a more rapid deepening of sea level pressure of the simulated storm is observed than with only radiosonde observations. The predicted track of the winter storm is also altered, moving closer to the coast. Using NCEP multisensor hourly rainfall data as verification, the 36-h forecasts with both TOMS ozone and radiosonde observations outperform the radiosonde-only experiments. These results indicate that TOMS ozone data contain valuable meteorological information, which can be used to improve cyclone prediction.

Introduction

The Atlantic Ocean, Gulf of Mexico, Gulf Stream, Great Lakes, and Appalachian Mountains play prominent roles in the evolution of winter weather systems in the eastern region of the United States. These storms can bring snowfall, freezing precipitation, heavy rains, high winds, coastal flooding, and strong convection. The prediction of East Coast winter storms has been a major challenge to numerical weather prediction (NWP; Uccellini et al. 1984; Maglaras et al. 1995).

NWP is an initial value problem that is very difficult for mesoscale prediction because of the rapid nonlinear growth of errors in the initial conditions (ICs; Zhang et al. 2002). Current ICs for mesoscale models are mostly based on large-scale analysis, radiosonde data, and traditional initialization procedures, and are not sufficient for mesoscale weather research simulations. Satellites carrying the National Aeronautics and Space Administration (NASA) Total Ozone Mapping Spectrometer (TOMS) have provided high-resolution quantitative measurements of vertically integrated column ozone concentration (Heath et al. 1975). Meteorologists have noted that ozone distributions in the vicinity of the tropopause could be viewed as a surrogate for stratospheric potential vorticity (PV) and contain significant information regarding dynamical characteristics of synoptic-scale and mesoscale flow regimes and associated weather (Danielsen 1968; Shapiro et al. 1982). A strong correlation between vertically integrated ozone (O3) and upper-level troughs and ridges has been observed (Reed 1950; Davis et al. 1999). Is column ozone data useful to NWP?

Over the past several years, the potential applications of TOMS data to numerical weather analysis and forecasts have been explored at various research and operational centers. Quantitative retrieval of synoptic-scale winds using a piecewise PV inversion technique was shown by Davis et al. (1999). Their studies used the correlation relationship between total ozone and model PV quantities. Other applications include Reed (1950) who showed that the day-to-day changes in total ozone are well related with synoptic-scale horizontal advection and vertical motion. Shapiro (1980) suggested that temporal changes of ozone and condensation nuclei through a vertical flux divergence explain the mechanism for the stratospheric–tropospheric exchange in the vicinity of jet stream frontal zones associated with tropopause folds. Interannual variations of total ozone were found to operate coherently with variations of upwelling planetary wave activity from the troposphere (Fusco and Salby 1999). Furthermore, Ravetta and Ancellet (2000) succeeded in identifying dynamical processes at the tropopause during the decay of a cutoff low by comparing observed ozone and potential vorticity data.

Four-dimensional variational data assimilation (4DVAR) has proved to be a very useful method to assimilate indirect and irregular observations for mesoscale predictions, such as rainfall (Zupanski and Mesinger 1995; Zou and Kuo 1996), satellite brightness temperature (Zou et al. 2001), and water vapor wind vectors (Xiao et al. 2000). In this study, TOMS ozone and radiosonde data are used in a 4DVAR framework to assess TOMS ozone's added values to the prediction of a severe winter storm that occurred in January 2000 along the East Coast of the United States. This snow storm case is chosen to test the importance of a better description of tropopause folding that is contained in TOMS ozone on modeling the cyclone development. It also provides an opportunity to examine the impact of TOMS data over a relatively data-dense region. Most operational models failed to predict the heavy snow storm over the Washington, D.C., area. Studies on the role of moist processes and upper-level PV anomalies on the storm prediction can be found in Zhang et al. (2002). This study will investigate how the inclusion of TOMS ozone observations into a 4DVAR assimilation affects the 24- and 36-h forecasts of the Washington, D.C., snow storm.

In the following section, a brief overview of the winter storm is given. Experimental designs are described in section 3. Section 4 examines the relationship between TOMS ozone and mean potential vorticity (MPV). Numerical results from assimilation and prediction experiments are discussed in section 5. Summary and conclusions are given in the final section 6.

The winter snow storm system and observations

A synoptic overview

In the month of January 2000, temperatures across the United States were mild in comparison with climatological values (information was available online at http://lwf.ncdc.noaa.gov/oa/climate/research/2000/jan/jan00.html). However, a significant storm system began affecting the eastern seaboard on 22 January. First, a severe ice storm hit northern Georgia and portions of northwest South Carolina on 22–23 January. The storm initiated on 23 January with a surface low located in the southeastern United States near the coast of the Gulf of Mexico. Then, the low pressure system rapidly deepened and moved northward along the East Coast, with moist air from the ocean colliding with cold air over land. The geopotential height at 500 hPa and the sea level pressure (SLP) from the National Centers for Environmental Prediction (NCEP) analysis at 1200 UTC 25 January 2000 are shown in Figs. 1a and 1b. The flow feature is characterized by a surface low pressure system located on the East Coast of the United States, along with a midtropospheric trough. Heavy snow fell on the East Coast during 25–26 January 2000. Up to 17 in. of snow fell in parts of Virginia, and 20 in. over Raleigh, North Carolina. Ten inches accumulated outside Philadelphia, Pennsylvania. New York City, New York, measured 6 in., and up to 18 in. had fallen by the end of the storm in western Massachusetts.

TOMS ozone observations

We utilize Earth Probe (EP)/TOMS level-2 (along track, nongridded) data to provide the highest possible resolution of total ozone data. The EP/TOMS experiment provides measurements of the earth's total column ozone by measuring the backscattered earth radiance in six bands (308.60, 313.50, 317.50, 322.30, 331.20, and 360.40 nm). The measured radiances are normalized and compared with radiances derived by a radiative transfer calculation model, which generates a table of backscattered radiance as a function of total ozone and the conditions of the measurement. The total ozone value can then be derived by comparing measured radiances with calculated radiances and finding the ozone value that gives a calculated radiance equal to the measured radiance. TOMS ozone data are available once per day and are measured at local noon. The highest horizontal resolution (minimum distance between sample points in the direction of the scan) near nadir is 40 km (McPeters et al. 1998).

Uncertainties in the ozone values derived from TOMS measurements come from several possible sources: errors in the measurements of the radiances, errors of input physical quantities obtained from laboratory measurements, errors in the parameterization of atmospheric properties used as input to the radiative transfer computations, and limitations in the way the computations represent the physical processes in the atmosphere. Each of these sources of uncertainty can be manifested in a random error and an absolute error. For TOMS ozone data, the estimated absolute error is ±3%, and the random error is ±2%.

On 24 January 2000, observations from TOMS from 1200–1900 UTC covered most of the forecast domain (Fig. 2a). An ozone anomaly [TOMS > 360 Dobson units (DU)] was found near the coast of the Gulf of Mexico where the initial surface low pressure system associated with the Washington, D.C., snow storm is initially located (Fig. 2b). A maximum band of TOMS ozone was located along the northeast coast. The MPV field at 1200 UTC 24 January 2000 from the NCEP analysis (Fig. 2c) also have a maximum band extending from Alabama and Kentucky to Pennsylvania. Although the large-scale features between TOMS ozone and MPV are in general agreement, the strong ozone anomaly near the East Coast is not found in the analyzed field of MPV.

4DVAR formulation and experiment design

In 4DVAR experiments, the following cost function is minimized:
i1520-0450-42-6-797-e1
where x0 is the analysis vector on the analysis/forecast grid; xb is the NCEP analysis; d is a vector of observations consisting of either TOMS ozone, radiosondes, or data from both; R is the observation error covariance matrix; and B is the background error covariance matrix. A nonlinear operator M represents the numerical weather prediction model from initial time t0 to time ti, and H is the observation operator that transforms model variables to the observational quantities. The observation operator for TOMS ozone data assimilation first calculates MPV from wind and temperature fields and then transforms the MPV into the simulated ozone (see section 4 for details), while the observation operator for radiosonde data only interpolates the wind and temperature on the model grid points to the radiosonde locations. It is assumed that TOMS ozone observation errors are not correlated to radiosonde data errors and both errors are not spatially correlated. Through these assumptions, the observation error covariance matrix becomes diagonal. The observation error variances of radiosonde data are those from Parrish and Derber (1992). A constant value of 9.42 DU is used for the observation error variance of TOMS ozone, which is obtained by multiplying the percentage of the absolute observational error (3%) with an averaged TOMS ozone value (314 DU).

Three 4DVAR experiments are conducted to assess the impact of TOMS ozone data on the 36-h forecast of the Washington, D.C., snow storm. “BOTH” uses both the TOMS ozone and radiosonde observations, “RAD” includes only radiosonde observations, and “OZ” assimilates only TOMS ozone data. Forecasts without data assimilation (i.e., initialized with xb) are referred to as “CTRL.” A 7-h assimilation window from 1200 UTC 24 January to 1900 UTC 24 January 2000 is used for OZ and a 12-h assimilation window from 1200 UTC 24 January to 0000 UTC 25 January 2000 is used for RAD and BOTH. The fifth-generation Pennsylvania State University (PSU)–National Center for Atmospheric Research (NCAR) Mesoscale Model (MM5) 4DVAR system (Zou et al. 1997) is employed to conduct the experiments. The model version used for this study includes a cumulus parameterization scheme (Grell scheme), and the high-resolution planetary boundary layer parameterization scheme (Blackadar scheme). The land surface temperature is predicted using surface energy budget equations. For a more detailed description of MM5, see Dudhia (1993) and Grell et al. (1994). All data assimilation experiments are conducted over a horizontal grid mesh of 70 × 90 with 90-km grid resolution. There are a total of 27 layers in the vertical direction. These layers are defined at σ = 1.00, 0.99, 0.98, 0.96, 0.93, 0.89, 0.85, 0.81, 0.77, 0.73, 0.69, 0.65, 0.61, 0.57, 0.53, 0.49, 0.45, 0.41, 0.37, 0.33, 0.29, 0.25, 0.21, 0.17, 0.13, 0.09, 0.05, and 0.00.

Forecasts from the original analysis and the initial conditions generated by 4DVAR are carried out at 30 km (CTRL, OZ, RAD, and BOTH) on a single domain using the same physical options as those used in the data assimilation experiment. We have also carried out two 10-km experiments (RAD2 and BOTH2) nested in the 30-km domain, using the Goddard microphysical scheme. The vertical levels are kept the same in all data assimilation and forecast experiments. Forecasts at 10-km resolutions are carried out to examine the differences in the predictions of hourly precipitation amounts with and without TOMS ozone data.

Relationship between TOMS ozone and MPV

To incorporate TOMS ozone into a mesoscale model using a variational approach, an observation operator is needed to link the model variables to TOMS ozone data. In this study, an empirical linear correlation model between TOMS ozone and MPV is used. It is developed using a linear regression method, assuming a simple relationship between the vertically integrated ozone and MPV:
O3αβ,
where α and β are constants to be determined based on the statistics of TOMS ozone and MPV. Studies such as Allaart et al. (1993) and Beekmann et al. (1994) have quantified the correlation between vertically integrated ozone measured by satellite and MPV. MPV is calculated by
i1520-0450-42-6-797-e3
where ρ is the density, η = fk + ∇ × V is the three-dimensional absolute vorticity, θ is the potential temperature, P is the pressure, P2 = 100 hPa, P1 = 500 hPa, and ΔP = P1P2. The unit of MPV is the same as the PV unit (PVU) (1 PVU = 10−6 m2 K kg−1 s−1).

The constant coefficients α and β are determined statistically based on MPV calculated from either analyses or model forecasts and TOMS ozone observations during the 12-day period of 15–26 January 2000 over the United States. For example, Fig. 3 shows a scatter diagram between TOMS ozone and MPV using the 90-km analyses. A linear relationship between the two variables is observed.

In order to examine the sensitivity of the relationship between TOMS ozone and MPV to model resolution in space and time, three sets of model states are used: (i) 24-h mesoscale forecasts with hourly output at 30-km resolution (carried out 2 times per day during the 12-day period of 15–26 January 2000), (ii) 90-km MM5 standard analysis data at 12-h intervals, and (iii) 30-km MM5 standard analysis data at 12-h intervals. Estimates of α and β from the linear regression calculations using these three datasets are shown in Fig. 4. Differences in the values of α caused by the use of the three different MPV fields are larger than those of β. The value of α is slightly smaller using the 30-km forecasts than the analyses.

The linear relation between MPV and ozone is found to depend on the latitude in a similar fashion for all three datasets (Table 1 and Fig. 4), with a stronger correlation and larger value of α in middle latitudes (20°–60°N) than in the Tropics (0–20°N). The correlation between TOMS ozone and MPV is about 0.7–0.8 in middle latitudes and 0.2–0.3 in the Tropics. Similarly, α is greater than 20 between 20° and 60°N, but less than 10 between the equator and 20°N in all three cases.

Results in Figs. 3 and 4 are based on data from the 12-day period from 15 to 26 January 2000. The day-to-day variation of the correlation between TOMS ozone and MPV is shown in Fig. 5. The correlation drops below a value of 0.8 on 25–26 January 2000 when the Washington, D.C., snow storm occurred, clearly indicating either a forecast failure, a large uncertainty in the analyses at that time, a violation of the MPV–O3 correlation because of large latent heating, TOMS errors in the vicinity of significant deep cloud cover, or a combined result of several of these effects. It is noticed that MPV–O3 correlation is lowest for the 30-km analyses and highest for the 30-km forecasts. This may imply that the model forecasts were able to capture some useful information contained in the ozone observations that are not included in the 30-km analyses.

Numerical results

The control simulation based on the NCEP analysis data produced a forecast no better than operational forecasts. The 24-h forecasts from CTRL, BOTH, and RAD are shown in Fig. 6. The control simulation based on the NCEP analysis data failed to correctly simulate the location, intensity, and structure of the low pressure system associated with the snow storm. The low pressure system is elongated from the southwest to the northeast, with two split low centers. The simulated storm is much too weak and the track has an eastward bias compared to analysis data (Fig. 1a). Differences among CTRL, BOTH, and RAD are also shown in the predicted distribution of associated vertically integrated cloud water and rainwater. The simulated hydrometer fields are distributed more inland in both BOTH and RAD than in CTRL. The maximum rainwater in BOTH reached an area further south than RAD, and that of CTRL is found over the ocean and along the East Coast. Assimilation of TOMS ozone and radiosonde observations first adjusts the model initial conditions, which then modifies model forecasts. In the following, we examine initial perturbations produced by data assimilation and the perturbation growth during the following 24-h forecast period.

At 1200 UTC 24 January 2000, the tropopause at 360 K, which is marked by isentropic PV values of around 3.0 PVU, extends to as far south as Miami, Florida (Fig. 7a). The tongue of stratospheric air, characterized by a relative maxima in isentropic PV, extrudes equatorward along the East Coast and intensifies to more than 3 PVU after TOMS ozone assimilation (Figs. 7a,c). The relative minima in isentropic PV located over the West Coast is weakened in the meantime.

The 4DVAR analysis increments at the initial time (1200 UTC 24 January 2000) are shown in Fig. 7c for PV at 360 K and Fig. 8 for the 850-hPa temperature. The PV at 360 K increases by about 3 PVU along the south and East Coast of the United States when TOMS ozone data are assimilated. At 850 hPa, the 4DVAR analysis increments indicate a pair of negative and positive temperature anomalies being generated in RAD over Georgia and the coast where the initial low pressure system of the Washington, D.C., storm is located (Fig. 8b). Little change is introduced in the lower troposphere by only ozone assimilation (figure omitted). However, the addition of TOMS ozone observations to radiosonde observations slightly modifies the low-level temperature adjustments (Fig. 8c). These modifications in the upper levels from TOMS ozone, and in the lower levels from radiosonde data, shall favor the usual development of synoptic-scale systems (Bluestein 1993). Impacts of these types of initial adjustments to the prediction of a cyclone development were discussed in Zou et al. (1998). In the following, we examine the impacts of the initial PV adjustments generated by TOMS ozone assimilation to the prediction of the Washington, D.C., storm.

Although TOMS ozone assimilation is found to mainly change the PV distribution close to the tropopause at the initial time, the PV differences between BOTH and RAD propagate into lower layers. Figure 9 shows the cross sections of PV and potential temperature along a straight line from west to east cutting through the maxima of the isentropic PV at 360 K at 1800 UTC 24 January 2000 (6-h model forecast). The PV differences between BOTH and RAD is not confined to upper levels at these times because the tropopause undulates further into the lower levels and diabatic effects become evident at 6 h.

The initial perturbations introduced by data assimilation in RAD and BOTH experience rapid developments during the 24-h model forecasts. Figures 10 and 11 show the 24-h forecast increments of wind at 300 hPa (Fig. 10) and temperature at 850 hPa (Fig. 11) that are valid at 1200 UTC 25 January 2000. Although the initial perturbations are on the order of 1–2 m s−1 at 300 hPa, the maximum wind forecast increments are as large as 18 m s−1 and are found slightly north of the eastern Great Lakes area (Figs. 10b and 10c). Similarly, the low-level temperature perturbations in the experiments RAD and BOTH also experience a rapid development. A forecast increment of about 5 K is observed along the warm front located near the Washington, D.C., area at 1200 UTC 25 January 2000 (Figs. 11b and 11c). Differences between BOTH and RAD exist (Figs. 10d and 11d) but are smaller than the differences of any of them with respect to CTRL. The forecast increments from the OZ experiment remain at a similar magnitude, with the maximum differences between OZ and CTRL of 2–4 m s−1 for wind at 300 hPa, and less than 1 K for temperature at 850 hPa at 1200 UTC 25 January (figure omitted).

Figures 12, 13, and 14 provide a more detailed time evolution of the forecast differences between BOTH and CTRL. The wind (850 and 300 hPa) and temperature (850 hPa) differences between BOTH and CTRL are shown at 6-h intervals from 1800 UTC 24 January to 0600 UTC 25 January 2000. The initial 300 hPa 1–2 m s−1 wind speed perturbation associated with the storm increases to 5.5, 14.1, and 24.6 m s−1 at 1800 UTC 24 January, 0000 UTC 25 January, and 0600 UTC 25 January 2000, respectively (Fig. 12). The low-level wind speed also experiences a rapid growth, although not as great as in the upper levels (Fig. 13). The northwest tilt of the low pressure system is observed during the entire 24-h forecast period. Negative temperature anomalies found over the Georgia, South Carolina, and Virginia coast at the initial time (see Fig. 8d) moved southeastward with increasing magnitude during the next 18 h (Fig. 14). Although located initially underneath the upstream trough at 500 hPa, the positive temperature anomaly becomes collocated with the surface low pressure system as the cyclone develops into a major storm.

The time series of the central sea level pressure from the NCEP operational surface analyses, CTRL, OZ, RAD, BOTH, and “MIXED” are shown in Fig. 15a, where the forecast experiment MIXED takes the initial conditions below 500 hPa from RAD and above 500 hPa from CTRL. The SLP in BOTH and RAD is deeper than that of OZ and CTRL. Interestingly, very little difference is found if the upper-level adjustments at the initial time in RAD are removed. The experiment MIXED produced a similar level of deepening as that of RAD, indicating a dominant role of the low-level adjustments in RAD. This may explain why the assimilation of only TOMS ozone data made almost no impact on the subsequent cyclone development. When the low-level temperature anomaly exists, the addition of TOMS ozone data, which modifies the upper-level PV anomaly, is able to bring a more rapid cyclone deepening. Increasing the horizontal resolution improves the storm intensity prediction. The 10-km simulations (RAD2 and BOTH2) capture not only the storm's deepening better, but also the weakening in the last 6 h.

Figure 15b shows the storm's track errors for each experiment from 1200 UTC 24 January to 1200 UTC 25 January 2000. The track error of the RAD simulated storm is much smaller than that of the CTRL storm. Adding TOMS ozone further reduces the track errors, especially at 12 and 24 h. After 6 h, the tracks of BOTH and RAD are always located to the west of the track from CTRL and OZ, that is, the simulated storm tracks in both RAD and BOTH are closer to the coast. Also, the storm in BOTH is located closer to Washington, D.C., than those in CTRL and OZ at 1200 UTC 25 January 2000.

The precipitation from these experiments is compared with NCEP multisensor hourly precipitation data (Fig. 16). The main difference between the 24-h prediction occurred over the Washington, D.C., area and to its south. The maximum precipitation in BOTH is about 60.6 mm, which is lower than that of RAD. This region is near the observed maximum precipitation area. The time evolution of the hourly precipitation at four selected stations is shown in Fig. 17. It is found that RAD, BOTH, RAD2, and BOTH2 capture the initiation of precipitation at all four stations very well, while CTRL and OZ have a systematic delay. The experiment BOTH (BOTH2) produced a more realistic precipitation amount at all times than RAD (RAD2). The experiment BOTH2 provides the best hourly precipitation prediction.

The accumulated hourly precipitation from all 15 stations located in North Carolina, Virginia, and Maryland are compared with the observed accumulated hourly precipitation. Fifteen stations are selected at which the total observed precipitation amount from 1200 UTC 24 January to 0000 UTC 26 January exceeded 20 mm. The mean bias and root-mean-square (rms) errors are shown in Fig. 18. The rms errors of the hourly precipitation (Fig. 18a) from BOTH (BOTH2) are smaller than RAD (RAD2). A slight delay in precipitation is observed in RAD2 around 6 h. We noticed that the rms errors from the 30-km-resolution forecasts (RAD and BOTH) are smaller than those from the 10-km-resolution forecasts (Fig. 18a). It is found that the model tends to overpredict the precipitation amount at the higher resolution (10 km) and underpredict the precipitation amount at the 30-km resolution (Fig. 18b). At the same resolution (30 or 10 km), BOTH outperforms RAD.

Summary and conclusions

Assimilation of TOMS ozone data modifies the initial conditions of wind and temperature mostly in the upper troposphere, and assimilation of radiosonde data changes the initial conditions mainly in the lower levels. The combined use of both radiosonde and TOMS ozone generates an adjustment in the initial conditions that affects more significantly the 24–36-h forecast of the Washington, D.C., snowstorm than assimilating only one type of data. Within 24 h of model forecasts, the initial perturbations of both wind and temperature, initially located over the southeast coast region, experience a rapid amplification in the two experiments, including radiosonde observations. The development of the storm is found to be much more sensitive to the low-level perturbations than to the upper-level perturbations. The largest forecast differences with and without data assimilation are found near the surface low pressure system. Such an error growth is not observed in other regions or in the OZ experiment, which lacks the low-level perturbations. Use of TOMS ozone shows more benefit when radiosonde observations are included than when only TOMS data are used. Several forecast features of the storm are better captured by the assimilation of both TOMS ozone and radiosonde observations than the assimilation with only one type of data, or without any assimilation at all. These features include the intensity of the system, the track, the shape of the surface low system, the geographical distribution of the vertically integrated hydrometers, and the hourly precipitation at various stations along the East Coast.

These experiments demonstrate that the difficulty in the forecast of the Washington, D.C., snowstorm case is partially related to the rapid growth of forecast errors. The differences in the initial conditions that are relatively small and are comparable to the expected analysis errors can amplify into much more significant forecast differences after 24 and 36 h as the storm moves northeastward.

This study only suggests that TOMS ozone contain valuable meteorological information in the upper troposphere and could be used to improve mesoscale prediction given a reasonably accurate low-level analysis or additional observations in the low troposphere. For other cases for which large sensitivity exists in the upper troposphere, TOMS data alone might have significant forecast impacts. We realize that more case studies are needed before drawing any general conclusion on the usefulness of assimilating TOMS ozone data into a mesoscale model to improve numerical weather prediction.

Acknowledgments

This research is supported by National Aeronautics Space Administration project NAG5-11067.

REFERENCES

  • Allaart, M. A. F., H. Kelder, and L. C. Heijboer. 1993. On the relation between ozone and potential vorticity. Geophys. Res. Lett. 20:811814.

    • Search Google Scholar
    • Export Citation
  • Beekmann, M., G. Ancellet, and G. Megie. 1994. Climatology of tropospheric ozone in southern Europe and its relation to potential vorticity. J. Geophys. Res. 99:1284112853.

    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B. 1993. CLASS for class. Bull. Amer. Meteor. Soc. 74:16971702.

  • Danielsen, E. F. 1968. Stratospheric–tropospheric exchange based on radioactivity, ozone and potential vorticity. J. Atmos. Sci. 25:502518.

    • Search Google Scholar
    • Export Citation
  • Davis, C., S. Low-Nam, M. A. Shapiro, X. Zou, and A. J. Krueger. 1999. Direct retrieval of wind from Total Ozone Mapping Spectrometer (TOMS) data: Examples from FASTEX. Quart. J. Roy. Meteor. Soc. 125:33753391.

    • Search Google Scholar
    • Export Citation
  • Dudhia, J. 1993. A nonhydrostatic version of the Penn State–NCAR Mesoscale Model: Validation tests and simulation of an Atlantic cyclone and cold front. Mon. Wea. Rev. 121:14931513.

    • Search Google Scholar
    • Export Citation
  • Fusco, A. C. and M. L. Salby. 1999. Interannual variations of total ozone and their relationship to variations of planetary wave activity. J. Climate 12:16191629.

    • Search Google Scholar
    • Export Citation
  • Grell, G. A., J. Dudhia, and D. R. Stauffer. 1994. A description of the Fifth-Generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 138 pp. [Available from UCAR Communications, P. O. Box 3000, Boulder, CO 80307.].

    • Search Google Scholar
    • Export Citation
  • Heath, D. F., A. J. Krueger, A. J. Roder, and B. D. Henderson. 1975. The solar scatter ultraviolet and total ozone mapping spectrometer (SBUV/TOMS) for Nimbus G. Opt. Eng. 14:323331.

    • Search Google Scholar
    • Export Citation
  • Maglaras, G. J., J. S. Waldstreicher, P. J. Kocin, A. F. Gigi, and R. A. Marine. 1995. Winter weather forecasting throughout the eastern United States. Part I: An overview. Wea. Forecasting 10:520.

    • Search Google Scholar
    • Export Citation
  • McPeters, R. D., P. K. Bhartia, A. J. Krueger, and J. R. Herman. 1998. Earth Probe Total Ozone Mapping Spectrometer (TOMS) data products user's guide. NASA Tech. Publ. 1998-206895, 70 pp.

    • Search Google Scholar
    • Export Citation
  • Parrish, D. F. and J. C. Derber. 1992. The National Meteorological Center's spectral statistical-interpolation analysis system. Mon. Wea. Rev. 120:17471763.

    • Search Google Scholar
    • Export Citation
  • Ravetta, F. and G. Ancellet. 2000. Identification of dynamical processes at the tropopause during the decay of a cutoff low using high-resolution airborne lidar ozone measurements. Mon. Wea. Rev. 128:32523267.

    • Search Google Scholar
    • Export Citation
  • Reed, R. J. 1950. The role of vertical motions in ozone–weather relationships. J. Meteor. 7:263267.

  • Shapiro, M. A. 1980. Turbulent mixing within tropopause folds as a mechanism for the exchange of chemical constituents between the stratosphere and troposphere. J. Atmos. Sci. 37:9941004.

    • Search Google Scholar
    • Export Citation
  • Shapiro, M. A., A. J. Krueger, and P. J. Kennedy. 1982. Nowcasting the position and intensity of jet streams using a satellite-borne total ozone mapping spectrometer. Nowcasting, K. Browning, Ed., Academic Press, 137–145.

    • Search Google Scholar
    • Export Citation
  • Uccellini, L. W., P. J. Kocin, R. A. Petersen, C. H. Wash, and K. F. Brill. 1984. The Presidents' Day cyclone of 18–19 February 1979: Synoptic overview and analysis of the subtropical jet streak influencing the pre-cyclogenetic period. Mon. Wea. Rev. 112:3155.

    • Search Google Scholar
    • Export Citation
  • Xiao, Q., X. Zou, and Y-H. Kuo. 2000. Incorporating the SSM/I-derived precipitable water and rainfall rate into a numerical model: A case study for the ERICA IOP-4 cyclone. Mon. Wea. Rev. 128:87108.

    • Search Google Scholar
    • Export Citation
  • Zhang, F., C. Snyder, and R. Rotunno. 2002. Mesoscale predictability of the “surprise” snowstorm of 24–25 January 2000. Mon. Wea. Rev. 130:16171632.

    • Search Google Scholar
    • Export Citation
  • Zou, X. and Y-H. Kuo. 1996. Rainfall assimilation through an optimal control of initial and boundary conditions in a limited-area mesoscale model. Mon. Wea. Rev. 124:28592882.

    • Search Google Scholar
    • Export Citation
  • Zou, X., F. Vandenberghe, M. Pondeca, and Y-H. Kuo. 1997. Introduction to adjoint techniques and the MM5 adjoint modeling system. NCAR Tech. Note TN-435-STR, 110 pp. [Available from NCAR, P. O. Box 3000, Boulder, CO 80307-3000.].

    • Search Google Scholar
    • Export Citation
  • Zou, X., Y-H. Kuo, and S. Low-Nam. 1998. Medium-range prediction of an extratropical oceanic cyclone: Impact of initial state. Mon. Wea. Rev. 126:27372763.

    • Search Google Scholar
    • Export Citation
  • Zou, X., Q. Xiao, A. E. Lipton, and G. D. Modica. 2001. A numerical study of the effect of GOES sounder cloud-cleared brightness temperatures on the prediction of Hurricane Felix. J. Appl. Meteor. 40:3455.

    • Search Google Scholar
    • Export Citation
  • Zupanski, D. and F. Mesinger. 1995. Four-dimensional assimilation of precipitation data. Mon. Wea. Rev. 123:11121126.

Fig. 1.
Fig. 1.

(a) The geopotential height (solid) and temperature (dashed) at 500 hPa and (b) the sea level pressure (solid) and surface temperature (dashed) at 1200 UTC 25 Jan 2000. Contour intervals are 50 m for the geopotential height, (a) 2.5 and (b) 5.0 K for the temperature, and 4 hPa for the sea level pressure

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 2.
Fig. 2.

(a) TOMS ozone observation locations and (b) observed TOMS ozone field during 1200–1900 UTC 24 Jan 2000, and SLP at 1200 UTC 24 Jan 2000, and (c) MPV based on the NCEP analysis and sea surface pressure at 1200 UTC 24 Jan 2000. Contour interval is (b) 20 DU, (c) 1 PVU, and (b), (c) 4 hPa for the sea level pressure

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 3.
Fig. 3.

Scatterplot of MPV derived from the 90-km analysis vs TOMS total ozone observations over the same domain as Fig. 2 for the period 15–26 Jan 2000

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 4.
Fig. 4.

Values of (a) α and (b) β assuming O3 = α(MPV) + β, where O3 represents TOMS ozone and MPV is calculated from the three datasets: 90-km analyses, 30-km analyses, and 30-km forecasts

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 5.
Fig. 5.

Time series of the correlation coefficient between TOMS ozone and MPV for the three datasets (90-km analyses, 30-km analyses, and 30-km forecasts)

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 6.
Fig. 6.

The sea level pressure (solid), vertically integrated cloud water (ICLW) (dotted), and vertically integrated rainwater (IRNW) (shaded) at 1200 UTC 25 Jan 2000 for (a) CTRL, (b) BOTH, and (c) RAD. Contour intervals for SLP are 4.0 hPa and for ICLW are 0.2 mm. Areas where IRNW is greater than 0.1 and 0.3 mm are shaded lightly and heavily, respectively

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 7.
Fig. 7.

The PV at 360 K at 1200 UTC 24 Jan 2000 from (a) BOTH and (b) RAD. (c) The differences of PV at 360 K between BOTH and RAD (BOTH − RAD). Contour intervals are (a), (b) 1 and (c) 0.5 PVU.

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 8.
Fig. 8.

(a) The temperature at 850 hPa at 1200 UTC 24 Jan 2000 from CTRL. (b) Differences of temperature between RAD and CTRL. (c) Differences of temperature between BOTH and CTRL. (d) Differences of temperature between BOTH and RAD. Contour intervals are (a) 2.5, (b), (c) 0.2, and (d) 0.1 K.

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 9.
Fig. 9.

(a) The PV at 360 K (solid lines) and SLP (dotted lines) from BOTH, (b) cross section of PV (thin lines) and potential temperature (thick lines) along the line AB indicated in (a), and (c) the same as (b) but for the PV differences between BOTH and RAD (thin lines) and potential temperature (thick lines) at 1800 UTC 24 Jan 2000

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 10.
Fig. 10.

(a) The wind speeds and wind vectors at 300 hPa at 1200 UTC 25 Jan 2000 from CTRL. (b) Differences of wind speed between RAD and CTRL. (c) Differences of wind speed between BOTH and CTRL. (d) Differences of wind speed between BOTH and RAD. Contour intervals are (a) 10 and (b)–(d) 2 m s−1

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 11.
Fig. 11.

Same as Fig. 10, but for the temperature at 850 hPa. Contour intervals are (a) 2.5, (b), (c) 0.5, and (d) 0.2 K.

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 12.
Fig. 12.

Wind differences between BOTH and CTRL at 300 hPa for (a) 1800 UTC 24 Jan, (b) 0000 UTC 25 Jan, and (c) 0600 UTC 25 Jan. Contour interval is 2.0 m s−1

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 13.
Fig. 13.

Same as Fig. 12, but for the wind at 850 hPa

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 14.
Fig. 14.

The temperature differences (thin line) between BOTH and CTRL at 850 hPa and SLP (thick line) from BOTH for (a) 1800 UTC 24 Jan, (b) 0000 UTC 25 Jan, and (c) 0600 UTC 25 Jan. Contour interval is 0.5 K for temperature differences and 4 hPa for the sea level pressure.

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 15.
Fig. 15.

(a) Variation of the central SLP of the storm and (b) the cyclone track errors from the 90-km forecast experiments

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 16.
Fig. 16.

The accumulated total precipitation from 1200 UTC 24 Jan to 1200 UTC 25 Jan 2000 from (a) observations, (b) CTRL, (c) RAD, and (d) BOTH. Contour intervals for the precipitation are (a) 5.0 and (b)–(d) 10.0 mm. The solid dot indicated by “DC” in (a) indicates the position of Washington, DC. Solid circles indicate the precipitation observation sites that are used in Fig. 17. Solid circles and triangles indicate the 15 stations used in the calculation in Fig. 18

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 17.
Fig. 17.

Simulated and observed hourly precipitation amounts from 1200 UTC 24 Jan to 0000 UTC 16 Jan 2000 at (a) Camp Pickett, VA, (b) Norfolk, VA, (c) Carthage Hatteras, NC, and (d) Laurinburg, NC

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Fig. 18.
Fig. 18.

(a) Rms errors and (b) mean bias of hourly precipitation at 15 stations over NC, VA, and MD for which the observed total precipitation from 1200 UTC 24 Jan to 0000 UTC 26 Jan 2000 exceeded 20 mm

Citation: Journal of Applied Meteorology 42, 6; 10.1175/1520-0450(2003)042<0797:ITOMIT>2.0.CO;2

Table 1.

Correlation coefficients between TOMS ozone and MPV for the three datasets

Table 1.

Corresponding author address: Prof. X. Zou, Dept. of Meteorology, The Florida State University, 404 Love Bldg., Tallahassee, FL 32306-4052. zou@met.fsu.edu

* The National Center for Atmospheric Research is sponsored by the National Science Foundation.

Save