DO4models- Dust Observations for models: Linking a new dust source-area data set to improved physically-based dust emission schemes in climate models

沙尘是地球陆地-大气-海洋-生物圈系统的重要组成部分，影响气候、海洋肥力、陆地植物群落和人类健康。风能够将大量的灰尘从地球表面带到大气层中。仅北非每年就排放5亿至10亿吨灰尘。为了预测未来的天气和气候，数值模式是我们进行这种预测的关键工具，它代表了灰尘排放，运输和沉积的重要关系。从模型中排除灰尘会导致大的局部和全局误差。准确的粉尘建模始于正确的排放模拟。这是至关重要的，因为源区模拟误差会导致当地气候动力学和不正确的沙尘输送误差。然而，世界上许多主要的沙尘源区都位于极其偏远的地方，没有关于沙尘排放或控制的地面数据。虽然最近在确定主要沙尘源方面取得了进展，例如通过卫星数据，但许多沙尘排放模型仍然非常简单，不受真实的观测数据的限制。为了在有助于规划决策的空间尺度上预测天气和气候，数值模型提高了分辨率，因此一些全球模型的分辨率接近1度，许多区域模型的分辨率超过0.2度。观察到的表征沙尘源区和确实存在的行为的少数数据集根本不支持这些模型运行的规模。因此，要评估或改进模型中的粉尘排放量是极其困难的，因此，对粉尘源区域的模拟非常不准确，并且将继续如此。我们建议通过开发第一个模型粉尘排放计划来解决这个问题，该计划是基于专门建立的观测数据集，这些数据集是故意构建的，以精确匹配区域气候模型的规模。我们建议这样做，首先使用高分辨率的卫星数据，以确定在世界许多地方发现的灰尘源区的特点领域内的灰尘的主要来源。然后，我们将使用最先进的现场设备，系统地调查控制粉尘排放的真实的过程，在模型网格箱规模，测量背景条件在很长一段时间，以及发生在沙尘暴的重要过程。因此，我们将测量和监测控制地面上灰尘对风的可用性的因素（侵蚀性），以及风移动沉积物的能力（侵蚀性），以创建有关源区的明确数据集，这些数据集可用于未来几年的模型开发。我们将能够第一次确定什么样的灰尘源数据（例如表面粗糙度，土壤湿度，阵风）导致观测约束模式排放方案的最大改进。我们还将能够说明，如果不收集实地数据，只使用遥感数据作为模型的输入，模拟会产生什么样的误差。这将为未来如何以及在何处花费时间和金钱来改进气候模型提供重要的指导，并为收集哪种实地数据最重要提供方向。气象局没有能力进行广泛的实地工作，以提供对模型开发至关重要的观测数据。我们的建议跨越了野外工作、地球观测和数值模拟之间的传统障碍。只有这样，才能在尘埃数值模拟方面取得突破。我们前所未有的现场观测，这是专门为数值模式的需要，将是一个重要的一步，迈向新一代的模式计划。

Dust is an important part of the Earth's land-atmosphere-ocean-biosphere system affecting climate, the fertility of oceans, plant communities on land, and human health. Wind is able to move vast amounts of dust over the Earth's surface and into the atmosphere. North Africa alone emits 500-1000 million tons of dust a year. To predict future weather and climate it is crucial that numerical models, our key tool for such prediction, represent the relationships important to the emission, transport and deposition of dust. Excluding dust from models leads to large local and global errors. Accurate modelling of dust begins with the correct simulation of emission. This is vital because source area simulation errors lead to errors in local climate dynamics and incorrect dust transport. However, many of the major dust source regions of the world are in extremely remote places for which there is no ground-based data on dust emission or its controls. Although recent advances have been made in identifying major dust sources, for example from satellite data, many models of dust emission are still very simple and are not constrained by real observed data. In the drive to predict weather and climate at spatial scales useful for planning decisions, numerical models have increased their resolution so that some global models run at near 1 degree and many regional models at better than 0.2 degree resolution. The few observed data sets characterising dust source areas and behaviour that do exist simply do not support the scale at which these models are being run. It is therefore extremely difficult to either evaluate or improve the dust emission component of models as things stand. Simulation of dust source areas is consequently very inaccurate and is set to remain so. We propose to address this problem by developing the first model dust emission scheme which is based on purpose built observed data sets that have been deliberately constructed to exactly match the scale of regional climate models. We propose to do this by first using high-resolution satellite data to identify key sources of dust within field areas that are characteristic of dust source areas found in many parts of the world. We will then use state-of-the-art field equipment to systematically investigate the real processes that control dust emission at the model grid box scale, measuring background conditions over a long period as well as the important processes that occur during dust storms. We will therefore measure and monitor both the factors that control the availability of dust to the wind on the ground (erodibility), and the ability of the wind to move that sediment (erosivity) to create this definitive data set on source regions that can be used in model development for years to come. We will be able to determine for the first time what kind of dust source data (e.g. surface roughness, soil moisture, wind gustiness) lead to the largest improvement in the observationally-constrained model emission scheme. We will also be able to say what errors result in simulations if no field data is collected and only remotely sensed data are used as inputs to the models. This will provide important guidance on how and where to spend time and money in the improvement of climate models in the future and also to provide direction on what kind of field data are most important to collect. The Met Office does not have the capacity to undertake the extensive fieldwork required to deliver the observational data that are critical to model development. Our proposal cuts across the traditional barriers between field work, Earth Observation and numerical modelling. It is only by doing so that breakthroughs in dust numerical modelling will be achieved. Our unprecedented field observations which are tailored to numerical model needs will be a significant step towards a new generation of model schemes.