SCSMEX NESA (Northern Enhanced Sounding Array) averaged data fields

Version 3b (released on 07 December 2001)

About the Dataset Files

Version 3b of this dataset averages fields over the NESA polygon using the version 3b of the gridded data product (V3bGD). This V3bGD dataset was created using sounding data that has undergone additional QC over what was used in version 3. For example, some unQCed data got into the V3 analysis and the location of Shiyan 3 was incorrect in V3 on 05 June 06 UTC. Also, a bug was fixed in an interpolation routine which interpolated the sonde data to equally spaced 25 mb levels. In the earlier versions this bug resulted in a data fall out at certain levels. In the all V3 level products, JMA wind data was used at Laoag, Philippines below 700 mb to minimize topographics flow effects due to flow over the mountain range east of Laoag and JMA analyses were used for all fields at the surface. Budget derived rainfall estimates were computed using JMA fluxes that were adjusted towards fluxes collected on the R/V Shiyan 3 and processed by Dr. Parson's group at NCAR. Suspicious features in the Q1 and Q2 fields are still present and and indicate that additional quality control of the sounding data is necessary. The data are contained in the following four files:
  • basic fields as a function of pressure, fields.nesa
  • advection fields as a function of pressure, advect.nesa
  • heat and moisture fields as a function of pressure, q1q2.nesa
  • surface evaporation and budget-derived precipitation rates, eopo.nesa
  • To access these data files, click on the above hyperlinked filenames.

    This figure shows an inventory of sounding data during the SCSMEX IOPs. A completely filled circle indicates that soundings were available 152 times out of a possible 152 (4 soundings per day for 38 days). The red circles indicate the nominal positions of the research vessels located within the enhanced sounding arrays. The grid over the NESA shows the analysis domain for our gridded data product.

    Data in the above files represent an average of gridded data fields over the SCSMEX Northern Enhanced Sounding Array (NESA), that is, the northern polygon in the above figure. Fields of height (z), zonal wind (u), meridional wind (v), temperature (T) and specific humidity (q) were objectively analyzed using multiquadric interpolation (Nuss and Titley 1994) over a rectangular grid slightly larger than the NESA at 25 mb resolution from 1000 to 25 mb and at 1 degree resolution in both the latitude and longitude directions (see grid in above figure). Surface pressure was also analyzed over this grid. These analyses were done for 244 periods (61 days at 00, 06, 12, and 18 UTC) during May and June of 1998.

    Data used in this analyses included quality-controlled sounding data from 53 sounding sites around and over the South China Sea and the JMA reanalysis in data sparse regions (i.e., at every 5 degree grid intersection no observations within a 3 degree radius). During SCSMEX the ships were on site during the following periods:

      IOP 1: 05/06/1998 to 05/25/1998 (20 days)
      IOP 2: 06/05/1998 to 06/22/1998 (18 days)
    However the sounding sites over the Northern South China Sea (the NESA pologon) took observations 4x/day for the period from 05/05/98 thru 06/20/98. We refer to this 47-day period as the SCSMEX IOP during which time the results should be of higher quality due to the enhanced nature of the observations.


    • For the surface analyses: data were assumed to be at the surface only if the height of the observations were within 10 m of the station height. Surface pressure represents pressure reduced to sea level.
    • Negative precipitation rates, found in the file "eopo.nesa" result from a combination of factors: sampling errors (i.e., the inadequacy of a sounding at a given time and location to represent conditions in a large area and a longer time period), data errors, errors introduced by the objective analyses, and errors in the estimation of surface evaporation.
    • Evaporation rates were obtained from the JMA reanalyis and were adjusted towards the point flux measurement from collected on the R/V Shiyan 3.
    • To keep updated on any future problems, issues and/or changes to the data, users should register with Paul Ciesielski at
    File "fields.nesa" contains 244 periods of six-hourly data. For each six hour period there are 42 lines of data.
    Line 1 contains: [year, month, day, hour]
                      written with format (4i3)
    Lines 2-42 contain: [p(mb), z(m), u(m/s), v(m/s), omega(mb/hr), T(C), 
                        theta(K), specific humidity(gr/kg) and 
                        divergence(1/s)*10e6] written with format (9f.2)
    Omega is computed using O'Brien's (1970) method with an isovalue divergence adjustment. Omega is set to zero at the surface and at 100 mb.

    File "advect.nesa" contains 244 periods of six-hourly data. For each six hour period there are 42 lines of data.

    Line 1 contains: [year, month, day, hour]
                     written with format (4i3)
    Lines 2-42 contain: [p(mb), hu(m/s**2), vu(m/s**2), hv(m/s**2), vv(m/s**2),
                              hT(C/s), vT(C/s), hq(gr/(gr*s)), vq(gr/(gr*s))]
                              written with format (f8.2,1p,8e11.3)
                              where hu - horizontal advection of u
                              where vu -   vertical advection of u
                              where hv - horizontal advection of v
                              where vv -   vertical advection of v
                              where hT - horizontal advection of T
                              where vT -   vertical advection of T
                              where hq - horizontal advection of q
                              where vq -   vertical advection of q
    These horizontal and vertical advection terms were computed using centered differences as follows:
    horizontal advection of "f": h(f) = u*df/dx + v*df/dy
                          where: dx = a cos(phi)*d(lambda)
                                 dy = a d(phi)
                                 phi    - latitiude
                                 lambda - longitude
    vertical advection of "f": v(f) = omega*df/dp
    File "q1q2.nesa" contains 244 periods of six hourly data. The computation of Q1 and Q2 were not done for the first and last periods of the IOP because time center differences were used in computing these fields and no grids were analyzed before and after the IOP. For each six hour period there are 42 lines of data.
    Line 1 contains: [year, month, day, hour]
                     written with format (4i3)
    Lines 2-42 contain: [p(mb), Q1(C/day), Q2(C/day)]
                              written with format (3f.2)
    The apparent heat source, Q1, and moisture sink, Q2 (Yanai et al. 1973) were computed as:
    Q1/cp = [dT/dt + h(T) + (p/po)**kappa * omega * d(theta)/dp]
    Q2/cp = -Lv/cp * [dq/dt + h(q) + v(q)]
          where dt = 12 hours
                po = 1000 mb
                cp = 1004
                Lv = 2.5e6
                 g = 9.8
    File "eopo.nesa" contains values of surface evaporation (e_0) and precipitation rates (p_0) for 244 six-hourly periods. The surface evaporation values represent the average from the JMA reanalyses. Once surface evaporation is known, rainfall rates can be computed from the moisture budget by integrating the equation for Q2 (shown above) from 1000 mb to 100 mb as follows:
    p_0 = e_0 + 1./(g*Lv) * [integral(Q2*dp) from 1000mb to 100mb]
    Line 1-244 contains: [year, month, day, hour, e_0(mm/day), p_0(mm/day)]
                         written with format (4i3,2f8.2)
    Nuss W.A. and D.W. Titley, 1994: Use of multiquadric interpolation for meteorological objective analysis. Mon. Wea. Rev., 22, 1611-1631.

    O'Brien J.J., 1970: Alternative solutions to the classical vertical velocity problem. J. Appl. Meteor., 9, 197-203.

    Yanai, M.S., S. Esbensen and J.H. Chu, 1973: Determination of bulk properties of tropical cloud clusters form large-scale heat and moisture budgets. J. Atmos. Sci., 30, 611-627.

    Please send any questions or comments to Paul Ciesielski,