TAMSAT rainfall climatology and time series for East Africa workshop

目前对非洲降雨气候变化的预测是模糊和不确定的。这在一定程度上是由于非洲的降雨气候监测不足。如果我们不确定当前的气候，我们怎么能提供可靠的预测未来会发生什么？因此，需要一套改进的非洲降雨数据集，这种数据集具有良好的空间覆盖范围，而且跨度足够长，可以探测长期趋势和估计变率。为了量化趋势和可变性，数据集还必须在时间上是同质的，也就是说，它必须基于一组一致的数据输入，并在整个时间段内使用相同的算法。这样的数据集对于分析过去的气候变率以及促进改进季节性降雨预报和气候变化预测将是非常宝贵的。这些改进对于更可靠的干旱预警，必要时人道主义组织更及时的反应，以及对整个非洲更好的长期经济规划都是迫切需要的。目前可用的雨量计数据过于稀疏和断断续续，无法提供所需的数据集。卫星方法提供良好的空间覆盖，但通常缺乏时间同质性，因为它们使用的是随时间变化的不同输入的混合物。唯一能够提供可靠估计、非洲全空间覆盖、长时间序列和时间同质性的卫星方法是雷丁大学开发的TAMSAT算法。非洲的时间序列是基于回溯使用1982年以来完整的气象卫星热红外档案，目前正在雷丁处理，将于2010年10月完成。总的目标是通过在非洲举办的一系列讲习班将时间序列和方法转让给最终用户。在非洲举办讲习班使成本最小化，使更多的非洲科学家能够参与其中，并提供了通过使用仅在国家气象部门内可用的额外雨量数据来增加校准的机会来改进降雨估计。谷歌资助的一个试点讲习班已经为埃塞俄比亚国家气象局运行，并且非常成功。在这一经验的基础上，在拟议的项目范围内，我们寻求于2011年1月在乌干达为东非地区的几个国家举办第二次讲习班。讲习班将向参与国提供其区域的完整时间序列，这些序列经当地校准后得到扩充，再加上软件继续添加到档案中。本着“知识交流”计划的精神，本地可用的测量数据所带来的时间序列的改善将反馈给气候研究界。除乌干达外，与会者还将包括来自苏丹和卢旺达气象部门的两名代表，以及来自SWALIM（索马里沃尔特和土地信息管理）的两名代表。讲习班的产出对索马里和卢旺达特别重要，因为这两个国家由于政治动荡而缺乏历史雨量数据。为支援今次及未来的工作坊，我们亦正申请拨款，以升级及简化现有的软件工具，以提供一个方便使用者的软件包，供工作坊使用，以进行雨量计质素控制、卫星估计校正及时间序列生成。从长远来看，我们的目标是继续与参与进一步分析和应用降雨数据集的科学家合作。将寻求机会资助在其他区域举办类似的讲习班。

Current predictions of rainfall climate change in Africa are vague and uncertain. This is in part due to the fact that African rainfall climate is poorly monitored. If we are not sure of the current climate, how can we provide reliable forecasts of what will happen in the future? There is therefore a need for an improved data set of African rainfall which has good spatial coverage and spans a sufficiently long time period to detect long term trends and estimate variability. In order to quantify trends and variability, the data set must also be temporally homogeneous, which is to say it must be based on a consistent set of data inputs and use the same algorithm for the whole time period. Such a data set would be invaluable for analysing past climate variability as well as facilitating improvements in seasonal rainfall forecasts and climate change predictions. These improvements are urgently needed for more reliable warnings of drought, more timely response of humanitarian organisations when necessary, and for better long term economic planning for Africa in general. Currently available rainfall data from gauges are too sparse and intermittent to provide the required data set. Satellite methods offer good spatial coverage but generally lack temporal homogeneity because they use a mixture of different inputs which vary in time. The only satellite approach which provides reliable estimates, full spatial coverage for Africa, long time series and temporal homogeneity is the TAMSAT algorithm, developed at the University of Reading. A time series for Africa based on retrospective use of the full Meteosat thermal infra red archive dating back to 1982 is now being processed at Reading and should be completed by October, 2010. The overall aim is to transfer the time series and methodology to end users via a series of workshops in Africa. Running workshops in Africa minimises costs, allows participation of more African scientists and provides the opportunity to improve the rainfall estimates by augmenting the calibration with additional raingauge data only available within the national meteorological services. A pilot workshop funded by Google has already been run for the Ethiopian National Meteorological Agency and was very successful. Building on this experience, within the proposed project we seek to run a second workshop in Uganda in January 2011 for several countries in the East African region. The workshop will provide participating countries with the full time series for their region augmented by local calibration, plus software to carry on adding to the archive. In the spirit of the 'Knowledge Exchange' programme, the improvements in time series made as a result of the locally available gauge data will be fed back to the climate research community. As well as Uganda, the participants will include 2 representatives from each of the Sudanese and Rwandan meteorological services and 2 representatives from SWALIM (Somali Walter and Land Information Management). Outputs from the workshop are particularly important for Somalia and Rwanda because of the lack of historic raingauge data due to political turbulence in these countries. To support this and future workshops, we are also requesting funding to upgrade and simplify existing software tools so as to provide a user-friendly software package which will be used within the workshops for raingauge quality control, satellite estimate calibration and time series generation. In the longer term, we aim to continue to collaborate with the scientists involved in further analysis and application of the rainfall data set. Opportunities will be sought for funding to run similar workshops in other regions.