HTML
-
The high-resolution (3 km, hourly) products of CMA-GD model (Fig. 1) were used in this paper from January 2018 to March 2021, with the initial times at 00:00 and 12:00 UTC and a valid forecast period of 0-24 h. The grid data were the nearest neighbor interpolated to 420 stations in Hunan Province (24-31° N, 108-115 ° E) in order to be consistent with observations. Moreover, the observational data was provided by the China Integrated Meteorological Information Sharing System (CIMISS) (Sun et al. [36]).
Figure 1. Domain of the CMA-GD model in Lambert map projection. The red line represents the location of Hunan Province.
Data division: All samples were divided into a training dataset, verification dataset, and test dataset (Table 1), with the verification set used to determine the integration solution and the test set used as an independent sample for forecasting. According to CMA's service requirements for short-time precipitation (T/CMSA 0013-2019), the hourly rainfall was divided into six grades at thresholds of 0.1, 2, 4, 8, and 20 mm h-1. Here a clear day occurred when the per-hour rainfall was less than 0.1 mm h-1, since 0.1 mm is the smallest detectable amount of rain gauge in China. And rainstorms occurred when the per-hour rainfall exceeded 8 mm h-1. Since the number of smallscale samples (> 4 mm) in the training set was greatly imbalance from that of the large-scale samples, the small-scale samples were under-sampled at a ratio of 1: 10, and the number of samples at each scale is ten times that of the severe rainstorm samples to obtain balanced samples.
Datasets Period < 0.1 0.1≤x < 2 2≤x < 4 4≤x < 8 8≤x < 20 > =20 Training (original) 2018.1—2019.12 11869522 1495536 143606 75625 36572 8114 Training (sampling) 2018.1—2019.12 81140 81140 81140 75625 36572 8114 Validation 2020.1—2021.3 random 50% 3416554 457119 49399 24087 10278 2107 Test 2020.1—2021.3 rest 50% 3352259 511314 56631 25761 11011 2332 Table 1. Sample sizes for the different grades of datasets from January 2018 to March 2021 (units: number).
-
Three techniques were used in this step to build the QPF correction models (Table 2), including deep neural network (DNN), frequency-matching (FM) and optimal threat score (OTS). Among them, the DNN technique aims to reduce the location and magnitude biases of rainfall. While FM and OTS can only eliminate the deviation in magnitude between QPF and the observation, they are unable to correct for pattern biases (Zhu and Luo [10]; Wu et al. [37]; Wu and Chen[38]). The thresholds of FM and OTS are 0.1, 2, 4, 8, and 20 mm.
Method Principle Effect Training window Deep neural network (DNN) Hinton [4] Predict classification of QPF 2018.1—2019.12 Frequency-matching (FM) Zhu and Luo [10] Adjust magnitude of QPF Past 30 days of the same year and the next 30 days of last year Optimal threat score (OTS) Wu et al. [37] Same as FM Same as FM Table 2. List of different techniques for improving quantitative precipitation forecast.
For DNN, the training set was used to obtain the classification of precipitation grades consistent with Table 1, and verification was performed with the verification set. Throughout the whole training process, DNN maintained the model with highest accuracy of classification in the verification set.
Hyperparameters: The Adam optimizer (Kingma and Ba [39]), cross-entropy loss function, and StepLR scheduler were used. The learning rate was set at 0.001 initially and was updated with the optimizer and scheduler. Batch normalization was used (Ioffe and Szegedy [40]). The number of epochs was set at 200. All of the stations were used here for unified modeling.
Selection of variables: From different levels and types of physical variables in CMA-GD, 51 physical variables (Table 3) related to precipitation and severe convection were selected. All variables were normalized to eliminate the dimensional difference between different variables before training.
Feature Description Rain Hourly precipitation CAPE Convective available potential energy CIN Convective inhibition K K index SI Showalter index IVT Integrated vapor transport IVTD Divergence of IVT U10/V10 10-meter U-wind and V-wind T2/Rh2 2-meter temperature and relative humidity Theta Equivalent potential temperature at 500/700/850/925/1000 hPa W Vertical velocity at 500/700/850/925/1000 hPa GH Geopotential height at 500/700/850/925/1000 hPa DIV/VOR Divergence and vorticity at 500/700/850/925/1000 hPa T/Td Temperature and dew point temperature at 500/700/850/925/1000 hPa Vapor Vapor flux at 500/700/850/925/1000 hPa Table 3. Physical variables used in DNN.
Mapping: To conduct a quantitative comparison with FM and OTS technologies, the DNN classification results should be transformed into QPF. The linearly mapping method from forecast probability to QPF was used here (Fig. 2). The rainfall is calculated as follows:
$$ y=\left\{\begin{array}{l} 0, k<1 \\ \mathrm{OBS}_{k}+\left(\mathrm{OBS}_{k+1}-\mathrm{OBS}_{k}\right) \frac{x_{k}-P_{\min }}{P_{\max }-P_{\min }}, k \in\{1, 2, 3, 4\} \\ \mathrm{OBS}_{5}+\frac{x_{k}-P_{\min }}{P_{\min }} \times \mathrm{OBS}_{5}, k \geqslant 5 \end{array}\right. $$ where x and y are the forecast probability and QPF for different precipitation grades, respectively; OBSk is the kth precipitation threshold, which is consistent with the thresholds of FM and OTS (0.1, 2, 4, 8, and 20); and P is the forecast probability set of different magnitudes for the calculated training dataset and validation dataset.
-
The optimal integration (OPI) method acquires the optimal forecast by integrating the best members in weak and severe precipitation forecasts. The steps are shown in Fig. 3. Firstly, a study area was selected. Among all the multi-method calibrated members, members A and B were assumed to have the optimal performance for weak precipitation and severe precipitation, respectively. Next, all stations in the study area were divided into two groups. The stations where members A or B showed severe precipitation were classified into G1, and the remaining stations were classified into G2. For stations in G1, the OPI model integrated the values of member B. For stations in G2, the OPI integrated the values of member A to obtain the integrated QPF. Finally, a new rule was introduced to avoid the problem of continuous precipitation under long-term statistical conditions. That is, for all severe precipitation stations in the integrated QPF, whether their precipitation grades in members A and B spanned more than two orders was checked. Stations with more than two orders were replaced with the average of members A and B.
Figure 3. Schematic diagrams of the principles of the OPI method. Assuming that members A and B have the best weak rainfall and severe rainfall among all the multi-method calibrated members, respectively. The black and red circles represent the severe rainfall stations of A and B, respectively, and the green part represents stations with weak rainfall. MA and MB indicate the precipitation magnitude of stations in A and B, respectively.
Then, a comparative experiment using a shorter sliding training window was conducted (Table 4). Specifically, the OPI experiment used the optimal severe rain member in the whole validation period as member B; the OPI-sliding experiment uses the sliding optimal rainstorm (≥8 mm) member in training window.
Experiments Member of weak rain Member of severe rain Training window OPI Overall optimum Overall optimum The whole period OPI-sliding Overall optimum Sliding optimum Same as FM Table 4. Design of the sensitivity experiments.
-
Two assessment indicators including threat score (TS) and clear-rainy TS were used to assess the QPF of different bias-corrected methods. Root-mean-square error (RMSE) and mean absolute error (MAE) were used to verify the performance of DNN. Table 5 shows the calculation formulas.
Indicator Expression Threat score (TS) $\frac{\mathrm{NA}}{\mathrm{NA}+\mathrm{NB}+\mathrm{NC}} $ Clear-rainy TS $\frac{\mathrm{NA}+\mathrm{ND}}{\mathrm{NA}+\mathrm{NB}+\mathrm{NC}+\mathrm{ND}} $ Root-mean-square error (RMSE) $\sqrt{\frac{1}{N} \sum\nolimits_{i=1}^{N}\left(F_{i}-O_{i}\right)^{2}} $ Mean absolute error (MAE) $\frac{1}{N} \sum\nolimits_{i=1}^{N}\left|F_{i}-O_{i}\right| $ Table 5. Assessment indicators used in the evaluation of QPF. NA, NB, NC, and ND represent the number of hits, misses, false alarms, and correct negatives, respectively. O and F stand for the observations and model results, respectively. i represents the ith sample, and N represents the number of valid samples.