Global atmospheric chemistry: Integrating over fractional cloud cover
Global atmospheric chemistry: Integrating over fractional cloud cover
复制标题
全球大气化学：积分部分云量
DOI：
10.1029/2006jd008007
复制
发表时间：
2007
影响因子：
--
通讯作者：
J. Penner
中科院分区：
文献类型：
--
作者：
J. Neu;M. Prather;J. Penner
[1]这里定义的一种新方法允许在网格正方形内对复杂云场进行光化学平均，并且可以很容易地在当前的全球模型中实现。根据观测或气象模式的诊断，有许多上覆云层的部分云覆盖可以在每个方格网格中产生数百到数千种不同的云廓线。我们定义了一种基于正交的方法，在此应用于在此云模式范围内平均光解速率的问题，这为在全球模型中模拟云内化学开辟了新的机会。我们选择了多达四个代表性的云剖面，优化了每个云的选择和权重，以最大限度地减少与整个云分布集的集成相比，光解速率的差异。为了实现我们的算法，我们将UCI fast-JX光解代码适应于T42L40分辨率下ECMWF预测模型的云统计数据。对于热带和中纬度地区，使用四种代表性大气的O3、NO2和NO3的网格平方平均光解速率与使用最大随机重叠方案得到的数百种或更多多云大气的平均速率相差最多3.2%的RMS。此外，自由对流层和边界层的偏置误差均小于1%。对于目前的近似方法，类似的误差显示为10-20%。正交法的误差小于选择最大随机重叠格式的不确定性。我们将该方法应用于不同云剖面上光化学的平均，并将其扩展到异质云化学。
[1] A new approach defined here allows for the averaging of photochemistry over complex cloud fields within a grid square and can be readily implemented in current global models. As diagnosed from observations or meteorological models, fractional cloud cover with many overlying cloud layers can generate hundreds to thousands of different cloud profiles per grid square. We define a quadrature-based method, applied here to the problem of averaging photolysis rates over this range of cloud patterns, which opens new opportunities for modeling in-cloud chemistry in global models. We select up to four representative cloud profiles, optimizing the selection and weighting of each to minimize the difference in photolysis rates when compared with the integration over the entire set of cloud distributions. To implement our algorithm, we adapt the UCI fast-JX photolysis code to the cloud statistics from the ECMWF forecast model at T42L40 resolution. For the tropics and midlatitudes, grid-square-averaged photolysis rates for O3, NO2, and NO3 using four representative atmospheres differ by at most 3.2% RMS from rates averaged over the hundreds or more cloudy atmospheres derived from a maximum-random overlap scheme. Further, bias errors in both the free troposphere and the boundary layer are less than 1%. Similar errors are shown to be 10–20% for current approximation methods. Errors in the quadrature method are less than the uncertainty in the choice of maximum-random overlap schemes. We apply the method to the averaging of photochemistry over different cloud profiles and outline extensions to heterogeneous cloud chemistry.