Martin Neussel (2008)

Die Nutzung von Langzeitarchiven globaler Fernerkundungsdaten (MODIS-EVI und -LCP, FAPAR) zur Charakterisierung von Umweltveränderungen im Tarim-Becken (NW-China)

One major challenge of the 21st century is the exposure to our global environment to the intensified usage of natural resources, triggered by population growth and prosperity, has led to an enormous grade of resource exploitation to quench the demands in a globalized world. This partially led to serious capacity shortages worldwide, e.g. in the energy sector. The availability of natural freshwater, an issue which scientists claim to have with the highest potential of conflict, is one major driving force for the sustainability of an environment. Eco-systematic processes of the earth's surface occur at different scales in time and intensity. The climate in particular determines the activity and seasonal development of vegetation. These dynamics are predominantly driven by temperature in the humid mid-latitudes and by the availability of water in semi-arid regions. Therefore, the 2000 UN Millennium Development Declaration called upon the international community to combat desertification and mitigate the effects of drought and floods, since human activities became the main modifying parameter for many ecosystems. Furthermore, the global community was committed to develop integrated water resources management and water efficiency plans by the year 2005 and to support developing countries and countries with economies in transition in their efforts to monitor and assess the quantity and quality of water resources (UN 2007). With the current human population growth, the expansion of irrigated agriculture has increased strongly worldwide (UNESCO 2006). Large ecosystems, like the Tarim Basin in NW-China, have for many years been largely affected by complex environmental problems related to anthropogenic water usage. Environmental changes such as the degradation of indigenous vegetation were directly correlated to poor water use efficiency management. The environmental problems reached an alarming extent during the 1990s, such that policy makers intended to solve the problem by conceiving huge projects like the World Bank's Tarim Basin Project II. The objective was a reduction of negative anthropogenic impacts to local ecosystem properties through various sanctions. Since then, almost two decades later, the all-important question is how these sophisticated improvements with a financial value of more than 4 Billion US$ took effect, as a large-scale model for an environmental restoration. Accounting for singe dynamics, e.g. seasonal dynamics of ecosystems and short to long-term changes in land cover composition requires multiple measurements in time and scale. With respect to the characterization of the earth's surface and its transformation due to global warming and human-induced global change, there is need for appropriate methods to determine the activity of vegetation and the change of land cover. Remote sensing established such methods and serviceable data sources for global environmental monitoring, using multi-spectral satellite sensors such as the Moderate Resolution Imaging Spectro-Radiometer (MODIS on the NASA/TERRA satellite platform). MODIS provides us with spectral information on a global scale and in a moderate spatial resolution. Since 2000, readily derived MODIS Land Cover Products on various environmental issues are available free of charge.

Since global monitoring has become commonplace in assessing global environmental processes, spectral information is derived by various satellite-systems. Usually, such spectral information has to be transformed first, before it is usable for an analysis. It is a characteristic feature of MODIS that scientific data on various topics (such as land cover, snow-cover, vegetation indices, etc.) is provided in addition to the raw spectral information. Such products have been produced for over 8 years, which offers their use for time-series analyses. This free archive of information on the global environment is unique in extent and availability, begging the question of the usefulness of MODIS sensor derived Land Cover Products for detecting environmental changes. Enhanced technical solutions and capacities provide us with sophisticated tools for a detailed analysis of MODIS products. As every multi-spectral sensor, the adaptability of MODIS spectral information has limitations. The moderate resolution of the sensor limits its prospects. Still, it offers the possibility of analysing large areas with a manageable amount of data.

A variety of factors disrupts the earth's ecosystem, of which the anthropogenic influence is on the track to ascendancy. Although the 'anthropogenic fingerprint' is not yet at its final stage, it has already changed landscapes worldwide. Especially freshwater and vegetation remain to be most important in this context, because they build the basis for human life. However, these resources are vulnerable to anthropogenic consumption and usage. Therefore, it is necessary to detect and analyse changes in the environment. This task can be achieved by evaluating vegetation changes and by utilizing the available MODIS products for instance. The question is how accurate and reliable the MODIS land cover products are in detecting vegetation changes. Although a number of studies have been conceived to utilize MODIS products for such purposes, only a few have been using them for detecting long-term changes on a large scale, e.g. Zhan (2002). One reason is the continuous enlargement of the MODIS data archive since 2000. Now that the MODIS sensor has collected data for more than 8 years the analysis of long-term time-series of the information enhanced. For this reason, a time-series has been constructed in order to detect environmental changes and to quantify the applicability of MODIS Land Cover Products in large hot-spot areas of anthropogenic influence such as the Tarim Basin in NW-China.

The study area ‘Tarim Basin’ has been selected because it represents a large fragile ecosystem that has been modified by intensive human activity. The core basin is represented by the arid and winter-cold Taklamakan Desert, which is surrounded by rivers that are mostly fed by glacial meltwater. In former times, such rivers used to supply large forests of indigenous vegetation through floods and groundwater. But since the 1950s’ large amounts of water that is diverted from the mountains has been used for irrigated agriculture. The consumption, also of groundwater, reached such enormous dimensions that the indigenous vegetation suffered from lack of water, and this caused sustained damage of the environment. In the 1990s’ the problem gained notoriety and decision-makers acknowledged need for action to counteract the negative impacts. In response, large projects with financial values up to a billion US$ have been launched by the World Bank, the Chinese government and by other institutions in order to remediate the basins ecosystem, and to improve the regional water management. By 2000, about 10 years had passed to revealing the outcome of the sophisticated projects. Although the project’s impacts have been partially analysed, a large-scale state description is still missing. Furthermore, constant changes or trends in the Tarim Basins environment have not yet been disclosed on a broad survey.

In order to detect such changes of the vegetation land cover, MODIS Land Cover Products have been used to create a time-series from March 2000 to February 2008 for almost the whole Tarim Basin. Differences in land use, land cover and changes within the period of time shall be separated and detected by using statistical tools. A time series-analysis based on Fourier algorithms should form the basis for a land cover classification. Stratified by this characterisation of vegetation properties, trend- and change-detections should reveal variances and a tendency of the on-going processes in the Tarim Basin. One major question is if the degradation of the indigenous 'Tugay'-vegetation will still continue, or if a stable or even positive trend is detectable with MODIS products. The dispersal of arable land within the study area and its degree of agricultural intensification, which is correlated to the availability and management of freshwater, will also be discussed. The focus will be set on the spatial extent as well as on the general condition of vegetation characteristics in the course of time.