Spatial patterns of high-elevation precipitation
observed through spaceborne radars
Masafumi Hirose1and Hatsuki Fujinami2
1: Meijo Univ, Japan 2: Nagoya Univ, Japan
Hight level 10
https://rain-clim.com
EGU21-15910, AS1.31
Upload date: March 31, 2021
High-altitude precipitation climates
↔Geographically inherent retrieval uncertainties
Topography [m] 10 km India Nepal BoB
AWS observation since 2016 (Fujinami et al. 2021, in rev, JGR-A. hereafter, F2021)
16 years of TRMM PR data, www.rain-clim.com
Long-term mean diurnal features estimated from spaceborne precipitation radar in high mountain regions ↔ in situ obs
86.49°E, 27.84°N
Fine-scale precipitation climatology has been examined on the basis of the long-term accumulation of spaceborne radar data (e.g., Hirose et al. 2017, JC). The first spaceborne precipitation radar, TRMM PR, has collected data for more than 16 years. As the number of samples increased, regional retrieval biases have become a vital interest for identifying geographic association of precipitation. This study is conducted to examine detection ability of precipitation variation over a complex terrain by using 1-km scale TRMM PR precipitation dataset with and without a correction for ground clutter interferences and to evaluate them with an experimental gauge observation installed at a high elevation in the Himalayas (F2021).
Target domain:
A part of Nepal Himalayas; Gaurishankar Conservation Area and the surrounding areas including Rolwaling valley where obs site was installed in 2016
Data:
Long-term mean precipitation dataset based onTRMM PR V8 1998-2013
+ AWS data at the Rolwaling valleyaround the Trambau Glacier terminus in the Nepal Himalayas.June-September, 2016-2018 (F2021)
Methods:
0.01° grid generation(Hirose and Okada 2018, JAMC), Low-level Precipitation Profile correction (Hirose et al., in rev, JMSJ)
Topography [m] 10 km [m] All NN SE
All: 1-49 bins, NN: Near nadir 21-23, 27-29 bins, SE: 1-2, 48-49 bins
0.71 deg. X 49 bins
CFB levels TRMM PR
Surface
CFB level
Nadir: 25th bin Precipitation rate [mm/h]
H ei gh t [km] CFB levels Surface Constant Ze≈ −4%/km in precip rate
Uncertainties related to different levels of clutter
free bottom (
CFB
) from surface
Original assumption
A retrieval uncertainty is found in incidence angle difference of precipitation in association with the ground clutter removal mask. The clutter free bottom (CFB) levels are generally high at its off nadir as compared to that at the near nadir. The CFB levels are relatively high in steep terrains. Below the CFB levels, precipitation is assumed to be decreased downward because of a pressure correction on the
terminal velocity for vertically constant attenuation-collected radar reflectivity factor (Ze). This algorithm assumption could be inappropriate for precipitation rate changes below approximately 2 km from the surface. Our 0.01° grid dataset shows that CFB levels could reach at nearly 3 km in the mountainous areas.
Precipitation Topography [m] [mm/day]
Ⓐ
10 kmJJAS
LPP-corrected surface precipitation
5 Precipitation rate [mm/h] H ei gh t [km] CFB levels Surface LPP DB based on near nadir statistics Input: land/ocean, stratiform/convective types, 0°C height, storm top height, vertical gradient of precipitation rate aloft
Output: LPP
[mm/day]
Effect of LPP correction
Ⓑ
100*(Ⓑ-Ⓐ)/Ⓐ
Correction based on low-level
precipitation profile (
LPP
) DB
Narrow valley
The high-resolution precipitation map clearly shows the spatial contrast of
0.1 deg. rain-clim.com
Diurnal features
Early morning Evening
Early afternoon Time of maximum precipitation around the Himalayas
Late morning
AWS → Twice-daily maxima of precipitation; afternoon and midnight. Precipitation frequency is approximately 50%. More than 80% of precipitation rate < 1 mm/h (F2021)
6
AWS
All season
0.01 deg.
One of our main target is evaluation of detection ability of local diurnal signature. The precedent researches based on the TRMM PR data reported afternoon peaks over most land, early afternoon peaks at the ridges or high mountain peaks, morning maxima in the foothill regions. Recently, F2021 found that bi-modal precipitation peaks appear at the high elevation area in the Nepal Himalayas. They reported that averaged precipitation rate is 0.3 mm/h < TRMM PR detectable threshold. This study compares the local features based on the 0.01° TRMM PR climatology and a referential gauge data.
Diurnal cycles at the Rolwaling AWS (F2021)
Time of max hourly rain
Precip1, 2: before and after the LPP correction
Topography
[m] 10 km
Contour: Altitude [m]
Ave within 1km (left) and 10km (right) from AWS (86.49E, 27.84N)
Fine-scale TRMM PR precipitation climatology around Trambau Glacier terminus in Nepal Himalaya
[mm
/d
ay]
JJAS
Corrected precipitation
Hourly precipitation and occurrence frequency
> 0.1 mm/h, JJAS 2016-2018
7
AWS
TRMM PR TRMM PR
In JJAS, the midnight precipitation peak and evening peak are significant in this lower and higher elevation areas, respectively. The hourly TRMM PR precipitation at the AWS site is calculated as a 3-hourly running average over adjacent pixels to reduce the sampling uncertainties found as a zig-zag temporal variation. The “precip2” in the line graph indicates a result from the LPP corrected precipitation data. The morning peak appears in precipitation amount but the occurrence
frequency of precipitation is relatively low. When data is averaged over areas within 10 km from the site, the morning peak disappears and the midnight and afternoon peaks become marked. The result is not fully consistent with the AWS observation data. The occurrence frequency and precipitation amount of TRMM PR precipitation is just one tenth and one fifth of the AWS data, respectively. The majority of
Summary and discussion
The long-term data can be used to extract geographic features of precipitation climate at a kilometer scale. Currently, 0.01° TRMM PR v.7 precipitation data is available via our web site [rain-clim.com].
The ground clutter mask could reduce surface precipitation by tens of % in the Himalaya areas where CFB levels are 2-3 km and shallow storms dominate. The low-level precipitation profile correction is to be incorporated in GPM DPR 07A algorithm as an experimental parameter. Even with this correction, precipitation amount and the occurrence frequency are much smaller than the ground observation data. As expected, sensitivity is a major issue over high elevation areas. Missing storms lower than the CFB level also result in significant underestimation (not shown).
The sampling of the short-term GPM DPR data is insufficient for discussions of km-scale precipitation, but detection ability of light precipitation is improved compared to TRMM PR (not shown). The upcoming DPR 07A algorithm will further improve detection of light precipitation by using 3D echo information. PR and DPR statistics differ by algorithms, but the combined data over a span of more than 2 decades will further update spatial and temporal coherency of precipitation in mountainous areas.
This study examined TRMM PR data properties for detecting km-scale precipitation features at high altitudes in the Himalayas. The valley precipitation becomes clear as the sample size increases. The profile assumption interfered by the ground clutter mask significantly reduces surface precipitation estimates. The spaceborne
precipitation radar captures only a part of whole precipitation systems due to sensitivity issues. Continuous algorithm updates of spaceborne radars and the combined use will produce better picture in near future.
References
Fujinami et al. (2021): Twice-daily monsoon precipitation maxima in the Himalayas driven by land surface effects. J. Geophys. Res.–Atmos., in revision Hirose and Okada (2018): A 0.01° Resolving TRMM PR Precipitation
Climatology, J. Appl. Meteor. Climatol., 57, 1645-1661.
Hirose et al. (2021): Refinement of surface precipitation estimates for the Dualfrequency Precipitation Radar on the GPM Core Observatory using near-nadir measurements, J. Meteor. Soc. Japan, in revision
Hirose et al. (2017): Spatial contrast of geographically induced rainfall observed by TRMM PR. J. Climate, 30, 4165-4184.
TRMM-PR captured Precipitation System Web site: https://www.rain-clim.com
Acknowledgements. This study was supported by the second EO RA of JAXA and
Backup slides
Precipitation Topography
[m]
[mm/day]
All: 1-49 bins, NN: 24-26 bin, SE: 1, 49 bins
Ⓐ
10 km
Correction by low-level precipitation profile (LPP) DB
All season
NN SE
LPP-corrected surface precipitation
11
[mm/day]
Effect of LPP correction