Source code for pyart.retrieve.qpe

"""
Functions for rainfall rate estimation.

"""

from warnings import warn

import numpy as np

from ..config import get_field_name, get_fillvalue, get_metadata
from .echo_class import get_freq_band


[docs]def est_rain_rate_zpoly(radar, refl_field=None, rr_field=None): """ Estimates rainfall rate from reflectivity using a polynomial Z-R relation developed at McGill University. Parameters ---------- radar : Radar Radar object. refl_field : str, optional Name of the reflectivity field to use. rr_field : str, optional Name of the rainfall rate field. Returns ------- rain : dict Field dictionary containing the rainfall rate. References ---------- Doelling et al. Systematic variations of Z–R-relationships from drop size distributions measured in northern Germany during seven years. 1998. Atmos. Ocean. Technol, 21, 1545-1556. Joss et al. Operational Use of Radar for Precipitation Measurements in Switzerland. 1998. Vdf Hochschulverlag AG ETH Zurich: 134. """ # parse the field parameters if refl_field is None: refl_field = get_field_name("reflectivity") if rr_field is None: rr_field = get_field_name("radar_estimated_rain_rate") radar.check_field_exists(refl_field) refl = radar.fields[refl_field]["data"] refl2 = refl * refl refl3 = refl * refl2 refl4 = refl * refl3 rr_data = np.ma.power( 10.0, -2.3 + 0.17 * refl - 5.1e-3 * refl2 + 9.8e-5 * refl3 - 6e-7 * refl4 ) rain = get_metadata(rr_field) rain["data"] = rr_data return rain
[docs]def est_rain_rate_z(radar, alpha=0.0376, beta=0.6112, refl_field=None, rr_field=None): """ Estimates rainfall rate from reflectivity using a power law. Parameters ---------- radar : Radar Radar object. alpha, beta : floats, optional Factor (alpha) and exponent (beta) of the power law. refl_field : str, optional Name of the reflectivity field to use. rr_field : str, optional Name of the rainfall rate field. Returns ------- rain : dict Field dictionary containing the rainfall rate. Reference --------- Fabry, Frédéric. Radar Meterology. 2015. Ch 9. pg 148-165. https://doi.org/10.1017/CBO9781107707405 """ # parse the field parameters if refl_field is None: refl_field = get_field_name("reflectivity") if rr_field is None: rr_field = get_field_name("radar_estimated_rain_rate") radar.check_field_exists(refl_field) refl = radar.fields[refl_field]["data"] rr_data = alpha * np.ma.power(np.ma.power(10.0, 0.1 * refl), beta) rain = get_metadata(rr_field) rain["data"] = rr_data return rain
[docs]def est_rain_rate_kdp(radar, alpha=None, beta=None, kdp_field=None, rr_field=None): """ Estimates rainfall rate from kdp using alpha power law. Parameters ---------- radar : Radar Radar object. alpha, beta : floats, optional Factor (alpha) and exponent (beta) of the power law. If not set the factors are going to be determined according to the radar frequency. kdp_field : str, optional Name of the specific differential phase field to use. rr_field : str, optional Name of the rainfall rate field. Returns ------- rain : dict Field dictionary containing the rainfall rate. Reference --------- Figueras et al. Long-term monitoring of French polarimetric radar data quality and evaluation of several polarimetric quantitative precipitation estimators in ideal conditions for operational implementation at C-band. Quarterly Journal of the Royal Meteorological Society. 2012. https://doi.org/10.1002/qj.1934 """ # select the coefficients as alpha function of frequency band if alpha is None or beta is None: # assign coefficients according to radar frequency if "frequency" in radar.instrument_parameters: alpha, beta = _get_coeff_rkdp( radar.instrument_parameters["frequency"]["data"][0] ) else: alpha, beta = _coeff_rkdp_table()["C"] warn( "Radar frequency unknown. " "Default coefficients for C band will be applied." ) # parse the field parameters if kdp_field is None: kdp_field = get_field_name("specific_differential_phase") if rr_field is None: rr_field = get_field_name("radar_estimated_rain_rate") radar.check_field_exists(kdp_field) kdp = radar.fields[kdp_field]["data"] kdp[kdp < 0] = 0.0 rr_data = alpha * np.ma.power(kdp, beta) rain = get_metadata(rr_field) rain["data"] = rr_data return rain
[docs]def est_rain_rate_a(radar, alpha=None, beta=None, a_field=None, rr_field=None): """ Estimates rainfall rate from specific attenuation using alpha power law. Parameters ---------- radar : Radar Radar object. alpha, beta : floats, optional Factor (alpha) and exponent (beta) of the power law. If not set the factors are going to be determined according to the radar frequency. a_field : str, optional Name of the specific attenuation field to use. rr_field : str, optional Name of the rainfall rate field. Returns ------- rain : dict Field dictionary containing the rainfall rate. References ---------- Diederich M., Ryzhkov A., Simmer C., Zhang P. and Tromel S., 2015: Use of Specific Attenuation for Rainfall Measurement at X-Band Radar Wavelenghts. Part I: Radar Calibration and Partial Beam Blockage Estimation. Journal of Hydrometeorology, 16, 487-502. Ryzhkov A., Diederich M., Zhang P. and Simmer C., 2014: Potential Utilization of Specific Attenuation for Rainfall Estimation, Mitigation of Partial Beam Blockage, and Radar Networking. Journal of Atmospheric and Oceanic Technology, 31, 599-619. """ # select the coefficients as alpha function of frequency band if alpha is None or beta is None: # assign coefficients according to radar frequency if "frequency" in radar.instrument_parameters: alpha, beta = _get_coeff_ra( radar.instrument_parameters["frequency"]["data"][0] ) else: alpha, beta = _coeff_ra_table()["C"] warn( "Radar frequency unknown. " "Default coefficients for C band will be applied." ) # parse the field parameters if a_field is None: a_field = get_field_name("specific_attenuation") if rr_field is None: rr_field = get_field_name("radar_estimated_rain_rate") radar.check_field_exists(a_field) att = radar.fields[a_field]["data"] rr_data = alpha * np.ma.power(att, beta) rain = get_metadata(rr_field) rain["data"] = rr_data return rain
[docs]def est_rain_rate_zkdp( radar, alphaz=0.0376, betaz=0.6112, alphakdp=None, betakdp=None, refl_field=None, kdp_field=None, rr_field=None, main_field=None, thresh=None, thresh_max=True, ): """ Estimates rainfall rate from a blending of power law r-kdp and r-z relations. Parameters ---------- radar : Radar Radar object. alphaz, betaz : floats, optional Factor (alpha) and exponent (beta) of the z-r power law. alphakdp, betakdp : floats, optional Factor (alpha) and exponent (beta) of the kdp-r power law. If not set the factors are going to be determined according to the radar frequency. refl_field : str, optional Name of the reflectivity field to use. kdp_field : str, optional Name of the specific differential phase field to use. rr_field : str, optional Name of the rainfall rate field. main_field : str, optional Name of the field that is going to act as main. Has to be either refl_field or kdp_field. Default is refl_field. thresh : float, optional Value of the threshold that determines when to use the secondary field. thresh_max : Bool, optional If true the main field is used up to the thresh value maximum. Otherwise the main field is not used below thresh value. Returns ------- rain_main : dict Field dictionary containing the rainfall rate. """ # parse the field parameters if refl_field is None: refl_field = get_field_name("reflectivity") if kdp_field is None: kdp_field = get_field_name("specific_differential_phase") if rr_field is None: rr_field = get_field_name("radar_estimated_rain_rate") rain_z = est_rain_rate_z( radar, alpha=alphaz, beta=betaz, refl_field=refl_field, rr_field=rr_field ) rain_kdp = est_rain_rate_kdp( radar, alpha=alphakdp, beta=betakdp, kdp_field=kdp_field, rr_field=rr_field ) if main_field == refl_field: rain_main = rain_z rain_secondary = rain_kdp elif main_field == kdp_field: rain_main = rain_kdp rain_secondary = rain_z elif main_field is None: main_field = refl_field rain_main = rain_z rain_secondary = rain_kdp else: main_field = refl_field rain_main = rain_z rain_secondary = rain_kdp thresh = 40.0 thresh_max = True warn( "Unknown main field. Using " + refl_field + " with threshold " + str(thresh) ) if thresh_max: is_secondary = rain_main["data"] > thresh else: is_secondary = rain_main["data"] < thresh rain_main["data"][is_secondary] = rain_secondary["data"][is_secondary] return rain_main
[docs]def est_rain_rate_za( radar, alphaz=0.0376, betaz=0.6112, alphaa=None, betaa=None, refl_field=None, a_field=None, rr_field=None, main_field=None, thresh=None, thresh_max=False, ): """ Estimates rainfall rate from a blending of power law r-alpha and r-z relations. Parameters ---------- radar : Radar Radar object alphaz, betaz : floats, optional Factor (alpha) and exponent (beta) of the z-r power law. alphaa,betaa : floats, optional Factor (alpha) and exponent (beta) of the a-r power law. If not set the factors are going to be determined according to the radar frequency. refl_field : str, optional Name of the reflectivity field to use. a_field : str, optional Name of the specific attenuation field to use. rr_field : str, optional Name of the rainfall rate field. main_field : str, optional Name of the field that is going to act as main. Has to be either refl_field or kdp_field. Default is refl_field. thresh : float, optional Value of the threshold that determines when to use the secondary field. thresh_max : Bool, optional If true the main field is used up to the thresh value maximum. Otherwise the main field is not used below thresh value. Returns ------- rain_main : dict Field dictionary containing the rainfall rate. """ # parse the field parameters if refl_field is None: refl_field = get_field_name("reflectivity") if a_field is None: a_field = get_field_name("specific_attenuation") if rr_field is None: rr_field = get_field_name("radar_estimated_rain_rate") rain_z = est_rain_rate_z( radar, alpha=alphaz, beta=betaz, refl_field=refl_field, rr_field=rr_field ) rain_a = est_rain_rate_a( radar, alpha=alphaa, beta=betaa, a_field=a_field, rr_field=rr_field ) if main_field == refl_field: rain_main = rain_z rain_secondary = rain_a elif main_field == a_field: rain_main = rain_a rain_secondary = rain_z elif main_field is None: main_field = a_field rain_main = rain_a rain_secondary = rain_z else: main_field = a_field rain_main = rain_a rain_secondary = rain_z thresh = 0.04 thresh_max = False warn("Unknown main field. Using " + a_field + " with threshold " + str(thresh)) if thresh_max: is_secondary = rain_main["data"] > thresh else: is_secondary = rain_main["data"] < thresh rain_main["data"][is_secondary] = rain_secondary["data"][is_secondary] return rain_main
[docs]def est_rain_rate_hydro( radar, alphazr=0.0376, betazr=0.6112, alphazs=0.1, betazs=0.5, alphaa=None, betaa=None, mp_factor=0.6, refl_field=None, a_field=None, hydro_field=None, rr_field=None, main_field=None, thresh=None, thresh_max=False, ): """ Estimates rainfall rate using different relations between R and the polarimetric variables depending on the hydrometeor type. Parameters ---------- radar : Radar Radar object. alphazr, betazr : floats, optional Factor (alpha) and exponent (beta) of the z-r power law for rain. alphazs, betazs : floats, optional Factor (alpha) and exponent (beta) of the z-s power law for snow. alphaa, betaa : floats, optional Factor (alpha) and exponent (beta) of the a-r power law. If not set the factors are going to be determined according to the radar frequency. mp_factor : float, optional Factor applied to z-r relation in the melting layer. refl_field : str, optional Name of the reflectivity field to use. a_field : str, optional Name of the specific attenuation field to use. hydro_field : str, optional Name of the hydrometeor classification field to use. rr_field : str, optional Name of the rainfall rate field. main_field : str, optional Name of the field that is going to act as main. Has to be either refl_field or kdp_field. Default is refl_field. thresh : float, optional Value of the threshold that determines when to use the secondary field. thresh_max : Bool, optional If true the main field is used up to the thresh value maximum. Otherwise the main field is not used below thresh value. Returns ------- rain : dict Field dictionary containing the rainfall rate. References ---------- Besic, N., Figueras i Ventura, J., Grazioli, J., Gabella, M., Germann, U., and Berne, A.: Hydrometeor classification through statistical clustering of polarimetric radar measurements: a semi-supervised approach, Atmos. Meas. Tech., 9, 4425–4445, https://doi.org/10.5194/amt-9-4425-2016, 2016. """ # parse the field parameters if refl_field is None: refl_field = get_field_name("reflectivity") if a_field is None: a_field = get_field_name("specific_attenuation") if hydro_field is None: hydro_field = get_field_name("radar_echo_classification") if rr_field is None: rr_field = get_field_name("radar_estimated_rain_rate") # extract fields and parameters from radar if hydro_field in radar.fields: hydroclass = radar.fields[hydro_field]["data"] else: raise KeyError("Field not available: " + hydro_field) # get the location of each hydrometeor class is_ds = hydroclass == 1 is_cr = hydroclass == 2 is_lr = hydroclass == 3 is_gr = hydroclass == 4 is_rn = hydroclass == 5 is_vi = hydroclass == 6 is_ws = hydroclass == 7 is_mh = hydroclass == 8 is_ih = hydroclass == 9 # compute z-r (in rain) z-r in snow and z-a relations rain_z = est_rain_rate_z( radar, alpha=alphazr, beta=betazr, refl_field=refl_field, rr_field=rr_field ) snow_z = est_rain_rate_z( radar, alpha=alphazs, beta=betazs, refl_field=refl_field, rr_field=rr_field ) rain_a = est_rain_rate_a( radar, alpha=alphaa, beta=betaa, a_field=a_field, rr_field=rr_field ) # initialize rainfall rate field rr_data = np.ma.zeros(hydroclass.shape, dtype="float32") rr_data[:] = np.ma.masked rr_data.set_fill_value(get_fillvalue()) # apply the relations for each type # solid phase rr_data[is_ds] = snow_z["data"][is_ds] rr_data[is_cr] = snow_z["data"][is_cr] rr_data[is_vi] = snow_z["data"][is_vi] rr_data[is_gr] = snow_z["data"][is_gr] rr_data[is_ih] = snow_z["data"][is_ih] # rain if main_field == refl_field: rain_main = rain_z rain_secondary = rain_a elif main_field == a_field: rain_main = rain_a rain_secondary = rain_z elif main_field is None: main_field = a_field rain_main = rain_a rain_secondary = rain_z else: main_field = a_field rain_main = rain_a rain_secondary = rain_z thresh = 0.04 thresh_max = False warn("Unknown main field. Using " + a_field + " with threshold " + str(thresh)) if thresh_max: is_secondary = rain_main["data"] > thresh else: is_secondary = rain_main["data"] < thresh rain_main["data"][is_secondary] = rain_secondary["data"][is_secondary] rr_data[is_lr] = rain_main["data"][is_lr] rr_data[is_rn] = rain_main["data"][is_rn] # mixed phase rr_data[is_ws] = mp_factor * rain_z["data"][is_ws] rr_data[is_mh] = mp_factor * rain_z["data"][is_mh] rain = get_metadata(rr_field) rain["data"] = rr_data return rain
def _get_coeff_rkdp(freq): """ Get the R(kdp) power law coefficients for a particular frequency. Parameters ---------- freq : float Radar frequency [Hz]. Returns ------- alpha, beta : floats The coefficient and exponent of the power law. """ coeff_rkdp_dict = _coeff_rkdp_table() freq_band = get_freq_band(freq) if (freq_band is not None) and (freq_band in coeff_rkdp_dict): return coeff_rkdp_dict[freq_band] if freq < 2e9: freq_band_aux = "S" elif freq > 12e9: freq_band_aux = "X" warn( "Radar frequency out of range. " + "Coefficients only applied to S, C or X band. " + freq_band_aux + " band coefficients will be used." ) return coeff_rkdp_dict[freq_band_aux] def _coeff_rkdp_table(): """ Defines the R(kdp) power law coefficients for each frequency band. Returns ------- coeff_rkdp_dict : dict A dictionary with the coefficients at each band. """ coeff_rkdp_dict = dict() # S band: Beard and Chuang coefficients coeff_rkdp_dict.update({"S": (50.70, 0.8500)}) # C band: Beard and Chuang coefficients coeff_rkdp_dict.update({"C": (29.70, 0.8500)}) # X band: Brandes coefficients coeff_rkdp_dict.update({"X": (15.81, 0.7992)}) return coeff_rkdp_dict def _get_coeff_ra(freq): """ Get the R(A) power law coefficients for a particular frequency. Parameters ---------- freq : float Radar frequency [Hz]. Returns ------- alpha, beta : floats The coefficient and exponent of the power law. """ coeff_ra_dict = _coeff_ra_table() freq_band = get_freq_band(freq) if (freq_band is not None) and (freq_band in coeff_ra_dict): return coeff_ra_dict[freq_band] if freq < 2e9: freq_band_aux = "S" elif freq > 12e9: freq_band_aux = "X" warn( "Radar frequency out of range. " + "Coefficients only applied to S, C or X band. " + freq_band_aux + " band coefficients will be used." ) return coeff_ra_dict[freq_band_aux] def _coeff_ra_table(): """ Defines the R(A) power law coefficients for each frequency band. Returns ------- coeff_ra_dict : dict A dictionary with the coefficients at each band. """ coeff_ra_dict = dict() # S band: at 10 deg C according to tables from Ryzhkov et al. 2014 coeff_ra_dict.update({"S": (3100.0, 1.03)}) # C band: at 10 deg C according to tables from Diederich et al. 2015 coeff_ra_dict.update({"C": (250.0, 0.91)}) # X band: at 10 deg C according to tables from Diederich et al. 2015 coeff_ra_dict.update({"X": (45.5, 0.83)}) return coeff_ra_dict
[docs]def ZtoR(radar, ref_field="reflectivity", a=300, b=1.4, save_name="NWS_primary_prate"): """ Convert reflectivity (dBZ) to precipitation rate (mm/hr) Author: Laura Tomkins Parameters ---------- radar : Radar Radar object used. ref_field : str Reflectivity field name to use to look up reflectivity data. In the radar object. Default field name is 'reflectivity'. Units are expected to be dBZ. a : float a value (coefficient) in the Z-R relationship b: float b value (exponent) in the Z-R relationship Returns ------- radar : Radar The radar object containing the precipitation rate field References ---------- American Meteorological Society, 2022: "Z-R relation". Glossary of Meteorology, https://glossary.ametsoc.org/wiki/Z-r_relation """ # get reflectivity data ref_data = radar.fields[ref_field]["data"] ref_data = np.ma.masked_invalid(ref_data) # convert to linear reflectivity ref_linear = 10 ** (ref_data / 10) precip_rate = (ref_linear / a) ** (1 / b) # create dictionary prate_dict = { "data": precip_rate, "standard_name": save_name, "long_name": f"{save_name} rescaled from linear reflectivity", "units": "mm/hr", "valid_min": 0, "valid_max": 10000, } # add field to radar object radar.add_field(save_name, prate_dict, replace_existing=True) return radar