@article{fdi:010062039,
  title = {{E}valuation of {X}-band polarimetric radar estimation of rainfall and rain drop size distribution parameters in {W}est {A}frica},
  author = {{K}offi, {A}. {K}. and {G}osset, {M}arielle and {Z}ahiri, {E}. {P}. and {O}chou, {A}. {D}. and {K}acou, {M}. and {C}azenave, {F}r{\'e}d{\'e}ric and {A}ssamoi, {P}.},
  editor = {},
  language = {{ENG}},
  abstract = {{A}s part of the {A}frican {M}onsoon {M}ultidisciplinary {A}nalysis ({AMMA}) field campaign an {X}-band dual-polarization {D}oppler radar was deployed in {B}enin, {W}est-{A}frica, in 2006 and 2007, together with a reinforced rain gauge network and several optical disdrometers. {B}ased on this data set, a comparative study of several rainfall estimators that use {X}-band polarimetric radar data is presented. {I}n tropical convective systems as encountered in {B}enin, microwave attenuation by rain is significant and quantitative precipitation estimation ({QPE}) at {X}-band is a challenge. {H}ere, several algorithms based on the combined use of reflectivity, differential reflectivity and differential phase shift are evaluated against rain gauges and disdrometers. {F}our rainfall estimators were tested on twelve rainy events: the use of attenuation corrected reflectivity only (estimator {R}({Z}({H}))), the use of the specific phase shift only {R}({K}-{DP}), the combination of specific phase shift and differential reflectivity {R}({K}-{DP},{Z}({DR})) and an estimator that uses three radar parameters {R}({Z}({H}), {Z}({DR}),{K}-{DP}). {T}he coefficients of the power law relationships between rain rate and radar variables were adjusted either based on disdrometer data and simulation, or on radar-gauges observations. {T}he three polarimetric based algorithms with coefficients predetermined on observations outperform the {R}({Z}({H})) estimator for rain rates above 10 mm/h which explain most of the rainfall in the studied region. {F}or the highest rain rates (above 30 mm/h) {R}({K}-{DP}) shows even better scores, and given its performances and its simplicity of implementation, is recommended. {T}he radar based retrieval of two parameters of the rain drop size distribution, the normalized intercept parameter {N}-w and the volumetric median diameter {D}-m was evaluated on four rainy days thanks to disdrometers. {T}he frequency distributions of the two parameters retrieved by the radar are very close to those observed with the disdrometer. {N}-w retrieval based on a combination of {Z}({H})-{K}-{DP}-{Z}({DR}) works well whatever the a priori assumption made on the drop shapes. {D}-m retrieval based on {Z}({DR}) alone performs well, but if satisfactory {Z}({DR}) measurements are not available, the combination {Z}({H})-{K}-{DP} provides satisfactory results for both {D}-m and {N}-w if an appropriate a priori assumption on drop shape is made.},
  keywords = {{C}onvective rainfall ; {P}olarimetric radar ; {T}ropical convection ; {Q}uantitative precipitation estimate ({QPE}) ; {D}rop size distribution ({DSD}) ; {BENIN}},
  booktitle = {},
  journal = {{A}tmospheric {R}esearch},
  volume = {143},
  numero = {},
  pages = {438--461},
  ISSN = {0169-8095},
  year = {2014},
  DOI = {10.1016/j.atmosres.2014.03.009},
  URL = {https://www.documentation.ird.fr/hor/fdi:010062039},
}