CW3E Publication Notice
The Impact of Dropsonde Data on the Performance of the NCEP Global Forecast System During the 2020 Atmospheric Rivers Observing Campaign. Part 1: Precipitation
October 20, 2022
Stephen J. Lord, UCAR Visiting Scientist at NOAA/NCEP/EMC, along with EMC collaborators Xingren Wu and Vijay Tallapragada, and CW3E Director F. Martin Ralph, has published the first of a two-part paper in Weather and Forecasting titled “The Impact of Dropsonde Data on the Performance of the NCEP Global Forecast System During the 2020 Atmospheric Rivers Observing Campaign. Part 1: Precipitation” (Lord et al. 2022). As part of CW3E’s 2019-2024 Strategic Plan to Support Atmospheric River (AR) Research and Applications, CW3E seeks to enhance global AR monitoring through a transformative modernization of atmospheric measurements over the Pacific and in the western United States. This study focuses on the improvements in precipitation forecasts produced by assimilating observations collected from dropsondes that were deployed during the 2020 AR Reconnaissance (ARR) Observing Campaign (OC). Specifically, this study is a detailed error analysis of key Intensive Observing Periods (IOPs) in the OC from 24 January (IOP-1) to 11 March 2020 (IOP-17).
The study uses analysis and forecast data from two NCEP Global Forecast System experiments, with (CTRL) and without (DENY) 628 dropsondes over the 17 IOPs covering the entire OC. The experiments are verified over two geographically separated domains, the Pacific Northwest and Northern California (PNNC) domain and the Southern California, Arizona and New Mexico (SCAN) domain. The dropsonde impact on precipitation forecasts is largely positive and appears driven by improvements to different model variables on a case-by-case basis. The results suggest that the important data gaps associated with ARs can be primarily addressed through targeted ARR field campaigns to provide vital, gap-filling observations needed to improve U.S. West Coast precipitation forecasts.
Several case studies illustrate the improvements and degradations in the CTRL forecast precipitation distributions relative to the DENY experiment when compared to the observed precipitation. For example, for the 24-h accumulated precipitation valid from 12 UTC 23–24 February (IOP-10), the observed CCPA (Climatologically Calibrated Precipitation Analysis, Version 4) domain-wide maximum precipitation was 71 mm and occurred over Washington State (Fig. 1a). The maximum amount forecasted in the CTRL experiment based on the 60–84 forecasts (initialized on 00 UTC 21 February 2020) is ~64 mm (i.e., 7 mm underestimation errors, Fig. 1b), while the maximum amount forecasted in the DENY experiment is ~56 mm (i.e., 15 mm underestimation errors, Fig. 1c). Therefore, the assimilation of dropsondes reduced 53% of the errors of the maximum precipitation amount in the DENY experiment. Differences between the DENY forecast rainfall and the CCPA verification (Fig. 2) were clearly reduced over the northern part of the domain in the CTRL relative to CCPA.
For selected heavy AR rainfall events (Table 1), 850 hPa geopotential height (Z850), wind speed (WSPD850) and specific humidity (SPCH850), impacts are correlated with impacts on the forecast Integrated (vertically) Vapor Transport (IVT). Results indicate IVT impacts are related most to impacts for wind speed (IOP-5), specific humidity (IOP-16) and geopotential height (IOP-17) in the different IOPs.
Figure 1: (a) 24-hour CCPA accumulated observed precipitation (mm) ending at 12 UTC 24 February. (b) CTRL 60-84 h forecast precipitation (mm) valid on 12 UTC 24 February. (c) As in Fig. 1b, except for the DENY forecast.
Figure 2: (a) Difference of the DENY 60-84 h forecast precipitation (mm) with the CCPA observed precipitation valid on 12 UTC 24 February. (b) As in Fig. 2a, except for the CTRL forecast.
Table 1. Correlations across 24-120 h forecast impacts between IVT and Z850, WSPD850 and SPCH850 for three cases.
|
IOP
|
IVT:Z850
|
IVT: WSPD850
|
IVT: SPCH850
|
|
5
|
0.112
|
0.710
|
0.670
|
|
16
|
0.537
|
0.389
|
0.725
|
|
17
|
0.613
|
0.370
|
0.101
|
|
Average
|
0.421
|
0.490
|
0.499
|
Lord, S., Wu, X. and Tallapragada, V. and Ralph, F. M., 2022a. The Impact of Dropsonde Data on the Performance of the NCEP Global Forecast System During the 2020 Atmospheric Rivers Observing Campaign. Part 1: Precipitation. Wea. Forecasting, https://doi.org/10.1175/WAF-D-22-0036.1
