Abstract:

Time-series analysis have been proven a powerful tool to analyse the climate system state and its natural and/or anthropogenic variability by means of historical data of the main meteorological variables (air temperature, relative humidity and precipitation), as well as terrestrial essential climate variables (ECVs) such as solar irradiation (hence, the photosynthetic active radiation, PAR), fAPAR (fraction of PAR), LAI (leaf area index) and land cover.

The time series (TS) are usually non-stationary, i.e., they present different frequency components, such as intra-annual variations (i.e., seasonal variability), short-term fluctuations resulting from disturbance events (i.e., forest fires) and long-term or inter-annual variations caused by external factors such as changing climatic conditions, land cover changes or land degradation. At this moment, global levels of archives of long TS with high temporal resolution are free available. These datasets includes meteorological variables at some ground stations such as mean, minimum and maximum temperature and precipitation as well as global maps of these variables at different temporal resolution (from daily to monthly temporal resolution).

Other type of very useful information is obtained from satellite data. Indeed, remote sensing has revealed as an excellent tool to obtain long series of Essential Climate Variables (ECVs), which can serve to map their spatial variability and to identify hot spot areas, as the areas showing the highest variability and considered as the most sensitive to climate variations. These data are mainly provided by European (MERIS/ENVISAT, SEVIRI/MSG, VEGETATION/SPOT) and America (TM/Landsat, MODIS & AQUA/TERRA, NOAA/AVHRR) satellites.

The analysis of long TS (three or four decades at least) is required in order to retrieve information about the state of the climate system and detect changes with statistical significance. Therefore, an appropriate and powerful tool is needed to retrieve information driven in these variables in a semi-automated way in order to evaluate not only by visual inspection the impact of climate change. In this sense, the free TimeStats tools have been used as an educational resource for a broad community of students in the University of Valencia. The TimeStats is programmed in the Interactive Data Language® (IDL) and freely distributed with the IDL virtual machine® for the analysis of multitemporal equidistant georeferenced remote sensing data archives, such as MODIS, AVHRR, MERIS and SPOT-Vegetation.

In this work, different activities and experiences of TS analysis application using the TimeStats for global climate change introduction in the educational area are summarized. These activities were aimed for students of different disciplines: Environmental management and land use, needing information on land cover type and land cover changes; Agricultural and forestry applications, requiring information on incoming/outgoing radiation and vegetation properties; Natural hazards management, which requires frequent observations of terrestrial surfaces. The teaching offers an overall view on the use of the TimeStats for TS analysis from regional to global conditions.

Keywords:

Time series, Climate system, remote sensing.