Qiang Dai

Appraisal of NLDAS-2 Multi-Model Simulated Soil Moistures for Hydrological Modelling

Soil moisture is a key variable in hydrological modelling, which could be estimated by land surface modelling. However the previous studies have focused on evaluating these soil moisture estimates by using point-based measurements, and there is a lack of attention for their appraisal over basin scales particularly for hydrological applications. In this study, we carry out for the first time, a detailed evaluation of five sources of soil moisture products (NLDAS-2 multi-model simulated soil moistures: Noah, VIC, Mosaic and SAC; and a ground observation), against a widely used hydrological model Xinanjiang (XAJ) as a benchmark at a U.S. basin. Generally speaking, all products have good agreements with the hydrological soil moisture simulation, with superior performance obtained from the SAC model and the VIC model. Furthermore, the results indicate that the in-situ measurements in deeper soil layer are still usable for hydrological applications. Nevertheless further improvement is still required on the definition of land surface model layer thicknesses and the related data fusion with the remotely sensed soil moisture. The potential usage of the NLDAS-2 soil moisture datasets in real-time flood forecasting is discussed.

Non-parametric rain/no rain screening method for satellite-borne passive microwave radiometers at 19–85 GHz channels with the Random Forests algorithm

This paper presents a novel non-parametric pattern recognition method to screen rain/no rain status for satellite-borne passive microwave radiometers in the 19–85 GHz channels. The method is based on randomized decision trees with bootstrap aggregation (Random Forests (RF) algorithm). It relies on pragmatic associations between the input features using Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) calibrated brightness temperatures and precipitation radar (PR) rain/no rain information as targets. Both these instruments are carried on board the TRMM satellite. In order to develop the method, first, the 10 most significant input features are selected by using feature importance criteria through out-of-bag (OOB) statistics from a total of 17 input features. The input features include the brightness temperatures, as well as some computed signatures – polarization differences (PD), polarization-corrected temperatures (PCT), and scattering indices (SI) at in the 19–85 GHz channels. The feature selection is carried out for different types of surface terrain (ocean, land, and coast), and the selected features are then used for final RF algorithm development. During the dichotomous statistical assessment of the method against the PR rain/no rain status as ‘truth’, the presented method produced reasonable threat scores of 0.50, 0.43, and 0.39, respectively, over ocean, land, and coast surface terrains. Furthermore, the results are compared with the dichotomous scores derived by the Goddard profiling algorithm (GPROF) and, remarkably, the RF-based method corroborated better statistical scores than that of the GPROF. The presented method does not rely on any a priori information and is applicable to other passive microwave radiometers at similar frequencies.

WRF Dynamical Downscaling and Bias Correction Schemes for NCEP Estimated Hydro-Meteorological Variables

Rainfall and Reference Evapotranspiration (ETo) are the most fundamental and significant variables in hydrological modelling. However, these variables are generally not available over ungauged catchments. ETo estimation usually needs measurements of weather variables such as wind speed, air temperature, solar radiation and dew point. After the development of reanalysis global datasets such as the National Centre for Environmental Prediction (NCEP) and high performance modelling framework Weather Research and Forecasting (WRF) model, it is now possible to estimate the rainfall and ETo for any coordinates. In this study, the WRF modelling system was employed to downscale the global NCEP reanalysis datasets over the Brue catchment, England, U.K. After downscaling, two statistical bias correction schemes were used, the first was based on sophisticated computing algorithms i.e., Relevance Vector Machine (RVM), while the second was based on the more simple Generalized Linear Model (GLM). The statistical performance indices for bias correction such as %Bias, index of agreement (d), Root Mean Square Error (RMSE), and Correlation (r) indicated that the RVM model, on the whole, displayed a more accomplished bias correction of the variability of rainfall and ETo in comparison to the GLM. The study provides important information on the performance of WRF derived hydro-meteorological variables using NCEP global reanalysis datasets and statistical bias correction schemes which can be used in numerous hydro-meteorological applications.

The impact of raindrop drift in a three-dimensional wind field on a radar–gauge rainfall comparison

There are many causes for the discrepancies between weather radar and rain gauges, and among these, displacement of raindrops due to wind drift – which is especially a problem with high-spatial resolution weather radar – is largely ignored in the published literature. This is mainly due to the lack of high-resolution three-dimensional wind fields and feasible treatment of the raindrop size distribution (DSD). In this study, a new systematic approach is proposed to explore the radar–gauge relationship under the wind influence. The mass-weighted mean diameter of raindrops is derived for each radar grid from the DSD data. The reanalysis project ERA-40 data of the European Centre for Medium-range Weather Forecasts (ECMWF) are used to drive the numerical weather research and forecasting (WRF) model to generate high-resolution hourly three-dimensional wind fields. Trajectories and displacements of raindrops are then computed using a three-dimensional motion equation from the given radar beam height to the ground surface. Based on the radar rainfall surface interpolated by the bicubic spline method, the correlation of the radar–gauge pairs is used to validate the results. A case study with 20 storm events in the Brue catchment in South West England is chosen to evaluate the proposed scheme. It has been found that when wind drift is taken into account, the correlation coefficient in hourly gauge–radar comparisons can be enhanced by up to 30% and the average correlation coefficient for an event can be improved by 10%. However, there are still some situations in which the scheme fails to work, indicating the complexity and uncertainties in tackling this challenging problem. Further studies are needed to explore why those cases cause problems to the scheme and how it could be improved to cope with them.

Tracking a tropical cyclone through WRF–ARW simulation and sensitivity of model physics

AbstractThe Weather Research and Forecasting (WRF) model’s Advanced Research WRF (ARW) dynamic solver is one of the most popular regional numerical weather prediction models being used by operational and research personnel. In this study, we simulate a tropical cyclone to reproduce the track direction and strength of the storm that formed at low latitudes in the West Pacific Ocean. The cyclone is known as “Haiyan” and assessed as category-5 equivalent super typhoon status due to its strong sustained winds and gusts, making it the strongest tropical cyclone in the region. We study the sensitivity of three different model physics options: the microphysics schemes, the planetary boundary layer schemes, and the impact of cumulus parameterization schemes. The realism of the cyclone simulation for different physics options is assessed through the comparison between the model outputs and the best track data, which are taken from the Japan Meteorological Agency. The experimental model simulations are carried out with two different global datasets: the ERA-Interim analysis from the European Centre for Medium Range Weather Forecasts and NCEP GFS forecast data, as initialization and boundary conditions. In addition, wind–pressure relationships are developed for different physics combination runs. Verification results associated with the model physics and boundary condition are discussed in this article. Overall, irrespective of the physics sensitivity, while the WRF simulation performs well in predicting the track propagation of the typhoon, substantial underestimation is seen in the intensity prediction.

Modelling radar-rainfall estimation uncertainties using elliptical and Archimedean copulas with different marginal distributions

Abstract Given that radar-based rainfall has been broadly applied in hydrological studies, quantitative modelling of its uncertainty is critically important, as the error of input rainfall is the main source of error in hydrological modelling. Using an ensemble of rainfall estimates is an elegant solution to characterize the uncertainty of radar-based rainfall and its spatial and temporal variability. This paper has fully formulated an ensemble generator for radar precipitation estimation based on the copula method. Each ensemble member is a probable realization that represents the unknown true rainfall field based on the distribution of radar rainfall (RR) error and its spatial error structure. An uncertainty model consisting of a deterministic component and a random error factor is presented based on the distribution of gauge rainfall conditioned on the radar rainfall (GR|RR). Two kinds of copulas (elliptical and Archimedean copulas) are introduced to generate random errors, which are imposed by the deterministic component. The elliptical copulas (e.g. Gaussian and t-copula) generate the random errors based on the multivariate distribution, typically of decomposition of the error correlation matrix using the LU decomposition algorithm. The Archimedean copulas (e.g. Clayton and Gumbel) utilize the conditional dependence between different radar pixels to obtain random errors. Based on those, a case application is carried out in the Brue catchment located in southwest England. The results show that the simulated uncertainty bands of rainfall encompass most of the reference raingauge measurements with good agreement between the simulated and observed spatial dependences. This indicates that the proposed scheme is a statistically reliable method in ensemble radar rainfall generation and is a useful tool for describing radar rainfall uncertainty. Editor D. Koutsoyiannis; Associate editor S. Grimaldi

Rain Rate Retrieval Algorithm for Conical-Scanning Microwave Imagers Aided by Random Forest, RReliefF, and Multivariate Adaptive Regression Splines (RAMARS)

This paper proposes a rain rate retrieval algorithm for conical-scanning microwave imagers (RAMARS), as an alternative to the NASA Goddard profiling (GPROF) algorithm, that does not rely on any a priori information. The fundamental basis of the RAMARS follows the concept of the GPROF algorithm, which means, being consistent with the Tropical Rainfall Measuring Mission (TRMM) precipitation radar rain rate observations, but independent of any auxiliary information. The RAMARS is built upon the combination of state-of-the-art machine learning and regression techniques, comprising of random forest algorithm, RReliefF, and multivariate adaptive regression splines. The RAMARS is applicable to both over ocean and land as well as coast surface terrains. It has been demonstrated that, when comparing with the TRMM Precipitation Radar observations, the performance of the RAMARS algorithm is comparable with the 2A12 GPROF algorithm. Furthermore, the RAMARS has been applied to two cyclonic cases, hurricane Sandy in 2012, and cyclone Mahasen in 2013, showing a very good capability to reproduce the structure and intensity of the cyclone fields. The RAMARS is highly flexible, because of its four processing components, making it extremely suitable for use to other passive microwave imagers in the global precipitation measurement (GPM) constellation.

Exploration of discrepancy between radar and gauge rainfall estimates driven by wind fields

Due to the fact that weather radar is prone to several sources of errors, it is acknowledged that adjustment against ground observations such as rain gauges is crucial for radar measurement. Spatial matching of precipitation patterns between radar and rain gauge is a significant premise in radar bias corrections. It is a conventional way to construct radar-gauge pairs based on their vertical locations. However, due to the wind effects, the raindrops observed by the radar do not always fall vertically to the ground, and the raindrops arriving at the ground may not all be caught by the rain gauge. This study proposes a fully formulated scheme to numerically simulate the movement of raindrops in a three-dimensional wind field in order to adjust the wind-induced errors. The Brue catchment (135 km2) in Southwest England covering 28 radar pixels and 49 rain gauges is an experimental catchment, where the radar central beam height varies between 500 and 700 m. The 20 typical events (with durations of 6–36 h) are chosen to assess the correlation between hourly radar and gauge rainfall surfaces. It is found that for most events, the improved rates of correlation coefficients are greater than 10%, and nearly half of the events increase by 20%. With the proposed method, except four events, all the event-averaged correlation values are greater than 0.5. This work is the first study to tackle both wind effects on radar and rain gauges, which could be considered as one of the essential components in processing radar observational data in its hydrometeorological applications.

Analysis of NDVI Data for Crop Identification and Yield Estimation

Crop yield estimation is of great importance to food security. Normalized Difference Vegetation Index (NDVI), as an effective crop monitoring tool, is extensively used in crop yield estimation. However, there are few studies focusing on the aspect of mixed crops grown together. In this study, a correlation-based approach for crop yield estimation is applied to three small counties (Jianshui, Luliang, and Qiubei) in the Nanpan River basin, Yunnan Province of China, and three main crops (paddy rice, winter wheat, and corn) in these areas are selected. Based on the correlation analysis between MODIS-NDVI data and crop yield, the crop planting areas as well as the best periods for a reliable estimation are identified. The best time is found approximately coinciding with the periods of heading, flowering, and filling of the crops. By Akaikes information criterion, the most fit regression models with extracted NDVI in the corresponding crop planting areas are determined. They work reasonably well in small regions, especially in the areas where crop types are unknown exactly. Further improvements to the regression models are possible by incorporating other physical factors such as soil types and geographical information.

CLOUDET: A Cloud Detection and Estimation Algorithm for Passive Microwave Imagers and Sounders Aided by Naïve Bayes Classifier and Multilayer Perceptron

CLOUDET, a cloud detection and estimation algorithm for passive microwave imagers and sounders is presented. CLOUDET is based on a naïve Bayes classifier and multilayer perceptron. It is applied to the special sensor microwave imager/sounder (SSMIS), and the ECMWF integrated forecast system (IFS) cloud liquid water information has been used to train the algorithm. CLOUDET is applicable to both ocean and land-surface types. CLOUDET has been developed and evaluated by employing two different groups of radiometric information, namely, the humidity channels near 183 GHz (humidity algorithm) and the window channels between 19 and 91 GHz (window algorithm). It has been revealed that both humidity and window algorithms can detect cloudy scenes over ocean at a confidence level of more than 90% (80% over land). The analysis indicates that the humidity algorithm has a better ability in detecting cloudy scenes over ocean than the window algorithm (CSI = 0.98 vs. CSI = 0.93). The opposite is true over land-surface type, revealing a CSI of 0.85 by humidity algorithm as opposed to CSI of 0.88 by window algorithm. The estimation of cloud by the CLOUDET algorithm has also been very promising during the validation effort. In particular, the correlation coefficient obtained over ocean through the use of the window algorithm is 0.70 (MAE 0.04 mm and RMSE 0.09 mm). The presented algorithm CLOUDET can be served as a stand-alone tool to reject and identify the cloudy scenes as well as to estimate the cloud liquid water path amount prior to assimilating the radiances into numerical weather prediction model.