Climatology-based Adjustments for Radar Rainfall in an OperaTional Setting: Adjustment factors for the Netherlands

This dataset contains gridded adjustment factors for correction of the Quantitative Precipitation Estimations (QPE) of the two operational C-band weather radars operated by the Royal Netherlands Meteorological Institute (KNMI). The factors are based on the CARROTS (Climatology-based Adjustments for Radar Rainfall in an OperaTional Setting) method, described in Imhoff et al. (2021). <br><br>The factors are available for every yearday (temporal resolution of one day) and are based on ten years (2009 - 2018) of radar and reference rainfall data, as distributed by KNMI. <br><br>For the derivation of the factors, both the operational radar QPE (https://doi.org/10.4121/uuid:05a7abc4-8f74-43f4-b8b1-7ed7f5629a01) and a reference rainfall dataset of KNMI (https://dataplatform.knmi.nl/catalog/datasets/index.html?x-dataset=rad_nl25_rac_mfbs_em_5min&amp;&amp;x-dataset-version=2.0) are used. The reference is not available in real time, but becomes available with a one to two month delay and was therefore available for this climatological factor derivation.<br>The derivation method was as follows per grid cell in the radar domain (Imhoff et al., 2021):<br>1. For every day in the period 2009--2018, an accumulation took place of all 5-min rainfall sums (of both the unadjusted radar QPE and the reference) within a moving window of 15 days prior to and 15 days after the day of interest. <br><br>2. For every yearday, the accumulations (per day) from the previous step were averaged over the ten years.<br>3. Gridded climatological adjustment factors (Fclim) were calculated per yearday as: Fclim(i,j) = RA(i,j) / RU(i,j). In this equation, RA(i,j) is the reference rainfall sum for the ten years and RU(i,j) the operationally available unadjusted radar QPE sum, based on the previous two steps, at grid cell (i, j).<br><br>For more details about the method, see Imhoff et al. (2021). For more information about the reference dataset, which consists of the radar QPE spatially adjusted with observations from 31 automatic and 325 manual rain gauges, see Overeem et al. (2009a,b).<br> <br><br>

本数据集包含用于校正荷兰皇家气象研究所（Royal Netherlands Meteorological Institute, KNMI）运行的两部业务化C波段天气雷达所产生的定量降水估测（Quantitative Precipitation Estimations, QPE）数据的网格化校正因子。该校正因子基于Imhoff等人（2021）所述的CARROTS方法（Climatology-based Adjustments for Radar Rainfall in an OperaTional Setting，基于气候学的业务场景雷达降雨调整法）计算得到。

该校正因子按年积日提供（时间分辨率为1天），其数据基础为KNMI发布的2009年至2018年共十年的雷达与参考降雨数据集。

本次校正因子的推导过程同时使用了业务化雷达QPE数据（https://doi.org/10.4121/uuid:05a7abc4-8f74-43f4-b8b1-7ed7f5629a01）与KNMI的参考降雨数据集（https://dataplatform.knmi.nl/catalog/datasets/index.html?x-dataset=rad_nl25_rac_mfbs_em_5min&&x-dataset-version=2.0）。该参考数据集并非实时可用，而是延迟1至2个月后发布，因此可用于本次气候学校正因子的推导计算。

针对雷达覆盖范围内的每个网格单元，其校正因子的推导步骤如下（Imhoff等人，2021）：
1. 在2009年至2018年的每一日，针对目标日前后各15天的移动窗口内所有5分钟降雨总量（包含未校正雷达QPE数据与参考降雨数据）进行累加求和。
2. 针对每个年积日，将上一步得到的单日累加值基于十年时段进行平均计算。
3. 按年积日计算网格化气候学校正因子（Fclim），计算公式为：Fclim(i,j) = RA(i,j) / RU(i,j)。其中，RA(i,j)为十年时段内网格单元(i,j)的参考降雨累加值，RU(i,j)为同网格单元基于前两步得到的业务可用未校正雷达QPE累加值。

如需了解该方法的更多细节，请参阅Imhoff等人（2021）的研究。关于该参考数据集的更多信息（该数据集为通过31个自动雨量计与325个人工雨量计观测数据对雷达QPE进行空间校正后的结果），请参阅Overeem等人（2009a、2009b）的相关文献。