A numerical method to generate high temporal resolution precipitation time series by combining weather radar measurements with a nowcast model

doi:10.1016/j.atmosres.2013.10.015

Atmospheric Research

Volume 138, 1 March 2014, Pages 1-12

https://doi.org/10.1016/j.atmosres.2013.10.015 Get rights and content

Highlights

•
Weather radar measurements are interpolated temporally using a numerical advection-model known from radar nowcasting.
•
The interpolation is able to reconstruct fine scale temporal variations of the rainfall.
•
This improves both the peak estimation and the accuracy of hour accumulated rain depths.
•
Improvements are verified by laser-based disdrometers.

Abstract

The topic of this paper is temporal interpolation of precipitation observed by weather radars. Precipitation measurements with high spatial and temporal resolution are, in general, desired for urban drainage applications. An advection-based interpolation method is developed which uses methods for vector field estimation already known from short-term weather radar nowcasting. However, instead of forecasting the weather radar rainfall, the proposed interpolation method exploits the advection of the rainfall in the interpolation. The interpolated rainfall fields are validated by measurements at ground level from laser disdrometers. The proposed interpolation method performs better when compared to traditional interpolation of weather radar rainfall where the radar observation is considered constant in time between measurements. It is demonstrated that the advection-based interpolation method reconstructs a significant part of the temporal variations of the rainfall. Moreover, the interpolated precipitation fields estimate accumulated rain depths more accurately.

Introduction

Precipitation measurements from weather radars are used as input to a wide variety of hydrological applications including flood risk management, hydropower production and urban drainage (e.g. Krajewski and Smith, 2002, Einfalt et al., 2004, Anagnostou et al., 2010, Tapiador et al., 2011). Measuring precipitation accurately can be a challenge due to the high variability in both time and space (Krajewski et al., 2003). Although rain gauges measure the temporal variability of the precipitation, the spatial resolution is limited unless a dense network is available. The required temporal and spatial sampling of the rainfall is achievable by weather radars. Weather radars are widely used by meteorological services worldwide, providing meteorologists with observations of rainfall intensity, Doppler velocities, hydrometeor classification etc.

The scanning strategy for weather radars is typically planned to facilitate meteorological services and may not be optimal for hydrological applications. The weather radars scan most of the atmosphere by changing the elevation angle of the antenna subsequently. Instead of scanning the whole atmosphere with volume scans, it might be more beneficial, from a hydrological point of view, to exchange the volume scan for a higher temporal resolution at the lowest several elevation angles. However, as most weather radars serve multiple purposes, changing the scanning strategy for runoff applications might not be feasible in the overall perspective.

Berne et al. (2004) and Einfalt et al. (2004) investigate the required rainfall resolution for urban drainage applications and suggest that resolution of 1–3 km in space and 1–5 min in time is sufficient for accurate runoff modelling of urban areas. Most modern weather radars fulfil this spatial requirement whereas the temporal requirement is seldom fully met. Typically, the temporal resolution is 5 to 15 min (Berne and Krajewski, 2013, OFCM, 2006, Piccolo and Chirico, 2005). In Denmark, the Danish Meteorological Institute (DMI) operates five meteorological C-band radars. The radars initiate a new scanning routine every 10 min with intermediate Doppler scans (Gill et al., 2006).

The radar reflectivity measurement represents an instantaneous spatial-distributed sample of the rainfall. However, if the precipitation-producing storm is moving, the rainfall might travel several kilometres between two successive radar scans. For a smaller urban catchment, this may result in an extreme situation where the radar observations show no precipitation over the catchment when a convective storm cell has in fact passed the area between the scans. Avoiding these temporal gaps in the radar observation is the key motivation for this study.

Previously, several authors (e.g. Fabry et al., 1994, Piccolo and Chirico, 2005, Shucksmith et al., 2011) have investigated the sampling error caused by the temporal gaps in radar rainfall. All three investigations found the uncertainty by comparing original high-resolution radar data with resampled coarser resolutions in both time and space. Shucksmith et al. (2011) found that the uncertainty on 10 min accumulations could be 70% or more if the estimate is conducted from a 10 min instant value and the precipitation field is assumed stationary within the sampling time interval. However, the uncertainty was significantly reduced if the storm movement and evolution were taken into account in the accumulation method (approx. errors on 25%). Fabry et al. (1994) introduced the advection-based accumulation methods whereas Piccolo and Chirico (2005) and Shucksmith et al. (2011) applied the concept on radar data with original finer resolution. Delobbe et al., 2006, Delobbe et al., 2008 evaluated 24 h accumulations with ground observations from rain gauges. However, it was concluded that the improvements were very limited even though the method succeeded in removing ripple effect due to the sampling problem.

The presented work is inspired by the rainfall accumulation methods of Fabry et al. (1994). However, instead of computing rainfall accumulations, the storm movement and evolution are used to generate higher temporal resolution of the radar rainfall measurements, which are more suitable for urban drainage runoff applications. The interpolation method makes use of the distributed vector field of the storm movement to produce temporal interpolations between the instantaneous radar scans. Moreover, different from Delobbe et al., 2006, Delobbe et al., 2008 the improvements of the advection-based interpolation are verified by ground observations from laser-based disdrometers.

Section snippets

Methods and materials

Taylor's hypothesis states that temporal variations in points are linked to the spatial variation through the convective velocity (Zawadzki, 1973, Einfalt and Lempio, 2009). This can be used to regenerate temporal variations from spatially distributed observations with a higher temporal resolution than originally sampled if the sampled spatial resolution is high and the convective velocity is known. In weather radar rainfall measurements, the spatial resolution is often much higher than the

Results

Fig. 6, Fig. 7, Fig. 8, Fig. 9 all illustrate selected rainfall events. The four events have been selected in order to illustrate how well the advection-based interpolation is performing under different meteorological conditions. Thus, the four events represent four different types of precipitation. Fig. 6 (Event 1) shows an event from 22 May 2011 where a frontal passage passed the area. The precipitation-producing storm system had a pattern with bands of precipitation perpendicular to the

Discussion

The results of this study clearly illustrate that the agreement between radar rainfall observations and ground observations is highly dependent on how the observations are compared. If observation with a high temporal resolution is evaluated, it is important that the observations are synchronized correctly as small time differences in the data may cause misleading interpretation of the radar performance.

The CAPPI or Pseudo-CAPPI radar data product (Constant Altitude Plane Position Indicator) is

Conclusions

The results demonstrate that the interpolation of the radar data is important especially if the data is to be used in hydrological applications, which require high resolutions in time and space. It is very important that the radar rainfall measurement is treated as an instantaneous measurement of the precipitation. If the radar estimated rainfall intensity is assumed constant in-between measurements, the correlation with the ground observations decreases significantly within several minutes.

Acknowledgements

The authors would like to thank the Danish Meteorological Institute (DMI). This work is part of the Storm and Wastewater Informatics project (SWI) partly financed by the Danish Agency for Science, Technology, and Innovation.

References (36)

M.N. Anagnostou et al.
Performance evaluation of high-resolution rainfall estimation by X-band dual-polarization radar for flash flood applications in mountainous basins
J. Hydrol.
(2010)
A. Berne et al.
Radar for hydrology: unfulfilled promise or unrecognized potential?
Adv. Water Resour.
(2013)
A. Berne et al.
Temporal and spatial resolution of rainfall measurements required for urban hydrology
J. Hydrol.
(2004)
T. Einfalt et al.
Towards a roadmap for use of radar rainfall data in urban drainage
J. Hydrol.
(2004)
F. Fabry et al.
High resolution rainfall measurements by radar for very small basins: the sampling problem reexamined
J. Hydrol.
(1994)
W.F. Krajewski et al.
Radar hydrology: rainfall estimation
Adv. Water Resour.
(2002)
S. Mecklenburg et al.
Improving the nowcasting of precipitation in an Alpine region with an enhanced radar echo tracking algorithm
J. Hydrol.
(2000)
J.E. Nielsen et al.
Calibrating LAWR weather radar using laser disdrometers
Atmos. Res.
(2013)
L.J. Battan
Radar Observation of the Atmosphere
(1973)
J. Choi
Study on Spatio-Temporal Properties of Rainfall
(2006)

L. Delobbe et al.

Combined use of radar and gauge observations for hydrological applications in the Walloon Region of Belgium

L. Delobbe et al.

Improvement of quantitative precipitation estimates in Belgium

T. Einfalt et al.

Some reflections on how to compare extreme point rainfall and extreme areal radar rainfall

Rainfall in the Urban Context: Forecasting, risk and, Climate Change, 10–13 December, 2009

(2009)

R.S. Gill et al.

The Danish Weather Radar Network

W.F. Krajewski et al.

An analysis of small-scale rainfall variability in different climatic regimes

Hydrol. Sci. J.

(2003)

L. Li et al.

Nowcasting of motion and growth of precipitation with radar over a complex orography

J. Appl. Meteorol.

(1995)

OFCM

Federal Meteorological Handbook No. 11 — Doppler Radar Meteorological Observations, Part C WSR-88D Products and Algorithms

(2006)

OTT Messtechnik

Operating Instructions: Present Weather Sensor Parsivel

(2009)

Cited by (34)

Merging weather radar data and opportunistic rainfall sensor data to enhance rainfall estimates
2024, Atmospheric Research
Accurate rainfall estimates with high spatiotemporal resolution are vital for urban drainage applications, and if adjusted to ground observations, weather radar data has potential for these applications. The density of meteorological rain gauges for adjustment is, however, often sparse and non-uniformly distributed in space. Opportunistic rainfall sensors provide an increased density of ground observations but often with reduced or unknown accuracy for each individual station. This paper demonstrates the merging of data from weather radar, personal weather stations, and commercial microwave links into an integrated rainfall product. Merging opportunistic rainfall estimates is shown to improve the accuracy of the opportunistic rainfall observations through a quality control algorithm. In this study, we show that the accuracy of rainfall estimates is significantly improved by merging opportunistic rainfall data and weather radar data when compared to the accuracy of each rainfall product without merging. Nash-Sutcliffe efficiency (NSE) values up to 0.88 are obtained for daily accumulated merged rainfall products, while the NSE-values of the individual rainfall products range from −7.44 to 0.65, and similar tendencies are observed for root mean squared error (RMSE) values. For merging weather radar and opportunistic rainfall data, a novel approach, i.e., “moving median bias adjustment” is presented. Applying this approach, a high-performing rainfall product is derived independently from conventional high-quality rain gauges, which in this study are only used for independent validation. In addition, it is demonstrated that accurate rainfall estimates can be obtained by sub-daily merging, underscoring the importance of merging in nowcasting and near real-time applications.
Comparing spatial metrics of extreme precipitation between data from rain gauges, weather radar and high-resolution climate model re-analyses
2022, Journal of Hydrology
The representation of extreme precipitation at small spatio-temporal scales is of major importance in urban hydrology. The present study compares point and radar observations to reanalyse climate model output data for a period of 14 years where there is full spatial and temporal overlap between datasets. The datasets are compared with respect to seasonality of occurrence, intensity levels and spatial structure of the extreme events. All datasets have similar seasonal distributions and comparable intensity levels. There are, however, clear differences in the spatial correlation structure of the extremes. Seemingly, the radar data is the best representation of a “real” spatial structure for extreme precipitation, even though challenges appear in data when moving far from the physical radar. The spatial correlation in point observations is a valid representation of the spatial structure of extreme precipitation. The convective-permitting climate model seems to represent the spatial structure of extreme precipitation much more realistically, compared to the coarser convective parameterized model. However, there is still room for improvement of the convective-permitting climate model for the shortest rainfall durations and smallest spatial scales in comparison with point and radar data.
Three-dimensional analysis of the initial stage of convective precipitation using an operational X-band polarimetric radar network
2019, Atmospheric Research
Citation Excerpt :
In the current study, we employed an algorithm for precipitation nowcasting. Different nowcasting algorithms have been proposed by many researchers; these algorithms can be roughly classified into two main categories: centroid methods (Crane, 1979; Bjerkaas and Forsyth, 1979; Browning et al., 1982; Austin and Bellon, 1982; Rosenfeld, 1987; Dixon and Wiener, 1993; Johnson et al., 1998; Handwerker, 2002; Han et al., 2009; Turso et al., 2013; Del Moral et al., 2018; Munoz et al., 2018) and cross-correlation methods (Bellon and Austin, 1978; Rinehart and Garvey, 1978; Smythe and Zrnic, 1983; Tuttle and Foote, 1990; Li et al., 1995; Lai, 1998; Mecklenburg et al., 2000; Li and Lai, 2004; Novak, 2007; Jung et al., 2011; Nielsen et al., 2014; Marra and Morin, 2018). Centroid methods identify and track the centroids of all individual storms.
In this study, we propose a new methodology for analysis of the initial stage of localized convective precipitation. The developed algorithm was used to create three-dimensional constant-altitude plan-position-indicator (3D CAPPI) data, which are high-spatiotemporal-resolution volumetric data, using two X-band polarimetric radars that are located in the Kanto region of Japan. Advection vectors are estimated by applying the normalized cross-correlation method to observed successive precipitation echoes on the polar coordinate system at each antenna tilt angle. The estimated advection vectors are expressed by a linear regression model that depends on time and tilt angle. In addition, 3D CAPPI data utilize the mosaic method to obtain further precipitation information from radar observations. The algorithm produces the detailed 3D structure of rapidly developing convective cells and provides quantitative information on convective precipitation, such as the echo top height, maximum reflectivity, and appearance time and height of each cell. The 3D CAPPI mosaic also clearly shows the back-building process at the initial stage of convective precipitation.
Enhanced object-based tracking algorithm for convective rain storms and cells
2018, Atmospheric Research
This paper proposes a new object-based storm tracking algorithm, based upon TITAN (Thunderstorm Identification, Tracking, Analysis and Nowcasting). TITAN is a widely-used convective storm tracking algorithm but has limitations in handling small-scale yet high-intensity storm entities due to its single-threshold identification approach. It also has difficulties to effectively track fast-moving storms because of the employed matching approach that largely relies on the overlapping areas between successive storm entities. To address these deficiencies, a number of modifications are proposed and tested in this paper. These include a two-stage multi-threshold storm identification, a new formulation for characterizing storm's physical features, and an enhanced matching technique in synergy with an optical-flow storm field tracker, as well as, according to these modifications, a more complex merging and splitting scheme. High-resolution (5-min and 529-m) radar reflectivity data for 18 storm events over Belgium are used to calibrate and evaluate the algorithm. The performance of the proposed algorithm is compared with that of the original TITAN. The results suggest that the proposed algorithm can better isolate and match convective rainfall entities, as well as to provide more reliable and detailed motion estimates. Furthermore, the improvement is found to be more significant for higher rainfall intensities. The new algorithm has the potential to serve as a basis for further applications, such as storm nowcasting and long-term stochastic spatial and temporal rainfall generation.
Applicability of open rainfall data to event-scale urban rainfall-runoff modelling
2017, Journal of Hydrology
Citation Excerpt :
The assumption of field stationarity between scans has been criticized as the storm field movement between scans is not accounted for (e.g. Fabry et al., 1994; Piccolo and Chirico, 2005; Shucksmith et al., 2011). To overcome this problem, creating new intermediate fields by means of advection interpolation between the precipitation fields has been proposed as a solution with encouraging results (Fabry et al., 1994; Nielsen et al., 2014; Wang et al., 2015). Accordingly, the 5 min data from Vantaa radar were advection interpolated to a temporal resolution of 1 min using an optical flow method of Farnebäck (2003).
Rainfall-runoff simulations in urban environments require meteorological input data with high temporal and spatial resolutions. The availability of precipitation data is constantly increasing due to the shift towards more open data sharing. However, the applicability of such data for urban runoff assessments is often unknown. Here, the feasibility of Finnish Meteorological Institute’s open rain gauge and open weather radar data as input sources was studied by conducting Storm Water Management Model simulations at a very small (33.5 ha) urban catchment in Helsinki, Finland. In addition to the open data sources, data were also available from two research gauges, one of them located on-site, and from a research radar. The results confirmed the importance of local precipitation measurements for urban rainfall-runoff simulations, implying the suitability of open gauge data to be largely dictated by the gauge’s distance from the catchment. Performance of open radar data with 5 min and 1 km² resolution was acceptable in terms of runoff reproduction, albeit peak flows were constantly and flow volumes often underestimated. Gauge adjustment and advection interpolation were found to improve the quality of the radar data, and at least gauge adjustment should be performed when open radar data are used. Finally, utilizing dual-polarization capabilities of radars has a potential to improve rainfall estimates for high intensity storms although more research is still needed.
Selection of meteorological parameters affecting rainfall estimation using neuro-fuzzy computing methodology
2016, Atmospheric Research
Rainfall is a complex atmospheric process that varies over time and space. Researchers have used various empirical and numerical methods to enhance estimation of rainfall intensity. We developed a novel prediction model in this study, with the emphasis on accuracy to identify the most significant meteorological parameters having effect on rainfall. For this, we used five input parameters: wet day frequency (d_wet), vapor pressure ( ${\overset{̅}{e}}_{a}$ ), and maximum and minimum air temperatures (T_max and T_min) as well as cloud cover (cc). The data were obtained from the Indian Meteorological Department for the Patna city, Bihar, India. Further, a type of soft-computing method, known as the adaptive-neuro-fuzzy inference system (ANFIS), was applied to the available data. In this respect, the observation data from 1901 to 2000 were employed for testing, validating, and estimating monthly rainfall via the simulated model. In addition, the ANFIS process for variable selection was implemented to detect the predominant variables affecting the rainfall prediction. Finally, the performance of the model was compared to other soft-computing approaches, including the artificial neural network (ANN), support vector machine (SVM), extreme learning machine (ELM), and genetic programming (GP). The results revealed that ANN, ELM, ANFIS, SVM, and GP had R² of 0.9531, 0.9572, 0.9764, 0.9525, and 0.9526, respectively. Therefore, we conclude that the ANFIS is the best method among all to predict monthly rainfall. Moreover, d_wet was found to be the most influential parameter for rainfall prediction, and the best predictor of accuracy. This study also identified sets of two and three meteorological parameters that show the best predictions.

View all citing articles on Scopus

View full text

A numerical method to generate high temporal resolution precipitation time series by combining weather radar measurements with a nowcast model

Highlights

Abstract

Introduction

Section snippets

Methods and materials

Results

Discussion

Conclusions

Acknowledgements

J. Hydrol.

Adv. Water Resour.

J. Hydrol.

J. Hydrol.

J. Hydrol.

Adv. Water Resour.

J. Hydrol.

Atmos. Res.

Radar Observation of the Atmosphere

Study on Spatio-Temporal Properties of Rainfall

Combined use of radar and gauge observations for hydrological applications in the Walloon Region of Belgium

Improvement of quantitative precipitation estimates in Belgium

Some reflections on how to compare extreme point rainfall and extreme areal radar rainfall

Rainfall in the Urban Context: Forecasting, risk and, Climate Change, 10–13 December, 2009

The Danish Weather Radar Network

An analysis of small-scale rainfall variability in different climatic regimes

Hydrol. Sci. J.

Nowcasting of motion and growth of precipitation with radar over a complex orography

J. Appl. Meteorol.

Federal Meteorological Handbook No. 11 — Doppler Radar Meteorological Observations, Part C WSR-88D Products and Algorithms

Operating Instructions: Present Weather Sensor Parsivel