NMQ Verification System (QVS) Tutorial

Figure 8. Mosaic 3D Derived Product main display.

Figure 9. Precipitation Flag (pcp_flag) display.

Figure 10. Hybrid Scan Reflectivity (hsr) display.

Figure 11. Difference field between unQC'd and QC'd composite reflectivity fields.

Figure 12.Scatter plot between unQC'd and QC'd composite reflectivity fields.

Tabs: Mosaic 3D Derived

This tab (Figure 8, at right) contains reflectivity products derived from the Mosaic 3D grid as follows:

unqc_cref - raw composite reflectivity

The composite reflectivity is the maximum reflectivity in each grid column in the 3-D mosaic. This cref has not been quality control edited to remove non-precipitating echoes. The unit of this field is dBZ.

cref - composite reflectivity

This cref has been edited with a neural network QC algorithm (QCNN; Lakshmanan 2007). The unit of this field is dBZ.

Lakshmanan, V., A. Fritz, T. Smith, K. Hondl, G. J. Stumpf, 2007: An automated technique to quality control radar reflectivity data. J. Appl. Meteor., 46, 288-305.

hgt_cref - height associated with the composite reflectivity

This variable is the height of the grid level where the column maximum reflectivity (composite reflectivity) is found. The unit is km above MSL.

pcp_flag - radar precipitation flag

The precipitation flag (example in Figure 9, at right, corresponds to HSR, seen in Figure 10) indicates the surface precipitation type and the radar observation representativeness at each grid point. The representativeness of radar observations with respect to surface precipitation estimation is highly correlated to the height of the radar observation and its proximity to the bright band layer (BBL) and to the surface. The top and bottom heights of the BBL are identified from the vertical profile of reflectivity (if it exists) at each radar site, and then objectively analyzed onto the NMQ grid using the RUC 0C height as the background.

Currently, each grid column in the 3-D mosaic grid is analyzed and classified as any one of the following:

-1: no radar coverage
0: no precipitation
1: surface precipitation is stratiform rain, and radar observation is below the bottom of the BBL
2: surface precipitation is stratiform rain, and radar observation is inside or above the BBL
3: surface precipitation is stratiform snow, and radar observation is below 1 km AGL
4: surface precipitation is stratiform snow, and radar observation is above 1 km AGL
5: mixed phase precipitation (place holder, not currently used)
6: convective precipitation
7: convective precipitation with potential hail contamination
8: orographically enhanced precipitation (place holder, not currently used)
9: tropical (warm cloud microphysics) precipitation

A grid column is classified as being convective if one of the following criteria is satisfied:

the reflectivity at any height outside bright band (if one exists) is greater than 50 dBZ, or
if the reflectivity at -10°C height or above is greater than 30 dBZ, or
if there is one or more lighting flashes in the bin location.

Using a VPR to determine if a grid column is tropical (reference Xu et al., 2007, .doc, 1.2 MB):

Using an hour-average VPR and the 0C level from the RUC, a VPR is searched downward from either the bottom of the brightband (as identified from the radar BB algorithm) or the 0C level.
A VPR that is nearly vertical or increases monotonically below the freezing level is assumed to be tropical (Houze, 1993). All grid points within that radar umbrella with echo > 30 dBZ are then designated as tropical and the tropical Z-R (see Tabs: Q2) is used.
Since storms may extend away from the radar umbrella into another radar’s domain that does not have enough data to compute a VPR, an algorithm assigns the tropical Z-R to echoes > 30 dBZ that are adjacent to the original tropical echo.

Temperature soundings are obtained from the RUC 20 km model analysis.

pcp_type - surface precipitation type

The 2-D surface precipitation type field is an integer-coded field that simply indicates the precipitation state at each grid cell. The precipitation state includes the following categories:

-1: no radar coverage/no data
0: no precipitation
1: liquid
3: frozen

etp18 - echo top (18 dBZ)

The echo top height is obtained by searching each grid column in the 3D reflectivity mosaic grid, from top to bottom, until the first reflectivity value greater than 18 dBZ is found. The height of the grid level where the reflectivity value was found is recorded as the ETP18 value for the given grid cell. The unit of the echo top is km above MSL.

shi - severe hail index

posh - probability of severe hail
mehs - maximum estimated hail size

The above three variables are hail products derived using the 3-D reflectivity mosaic grid and 3-D thermal field from the RUC 20km model analysis. The methodology is described in Witt et al. (1998). The severe hail index is dimensionless. The unit for the probability of severe hail is % and the unit for the maximum estimated hail size is mm.

Witt, A., M. D. Eilts, G. J. Stumpf, J. T. Johnson, E. D. Mitchell, and K. W. Thomas, 1998: An Enhanced Hail Detection Algorithm for the WSR-88D. Wea. Forecasting, 13, 286-303.

hsr - hybrid scan reflectivity

The hybrid scan reflectivity (Figure 10, above right) is a 2-D mosaic of single radar hybrid scan reflectivity fields. The single radar hybrid scan reflectivity is obtained by searching each grid column in the single radar 3D reflectivity Cartesian grid, from bottom to top, until the first non-missing reflectivity value is identified. The reflectivity value is then recorded as the single radar hybrid scan reflectivity (in dBZ) at the given grid cell, and the height of the grid cell is recorded as the single radar hybrid scan height.

hsrh - height associated with the hybrid scan reflectivity

This field is the height of the lowest non-missing single radar hybrid scan reflectivities at each NMQ grid cell. The unit of HSRH is meters AGL (above ground level).

vil - vertically integrated liquid (kg m^-2)

The VIL product is derived based on the procedures described in Greene and Clarke (1972).

Greene, D. R. and R. A. Clark, 1972: Vertically Integrated Liquid Water-A New Analysis Tool. Mon. Wea. Rev., 100, 548-552.

vilD - VIL density (g m^-3)

The VILD product is derived based on the procedures described in Amburn and Wolf (1997).

Amburn, S. A. and P. L.Wolf, 1997: VIL Density as a Hail Indicator. Wea. Forecasting, 12, 473-478.

Analysis Tools allow the following product comparisons:

difference between any two products with the same units (Figure 11, right)
scatter plot between any two products with the same units (Figure 12, right)