Jingjuan Liao

Lake variations in response to climate change in the Tibetan Plateau in the past 40 years

The Qinghai-Tibetan Plateau plays an important role in global climate and environmental change and holds the largest lake area in China, with a total surface area of 36,900 km2. The expansion and shrinkage of these lakes are critical to the water cycle and ecological and environmental systems across the plateau. In this paper, surface areas of major lakes within the plateau were extracted based on a topographic map from 1970, and Landsat MSS, TM and ETM+ satellite images from the 1970s to 2008. Then, a multivariate correlation analysis was conducted to examine the relationship between the changes in lake surface areas and the changes in climatic variables including temperature, precipitation, evaporation, and sunshine duration. Initial results suggest that the variations in lake surface areas within the plateau are closely related to the warming, humidified climate transition in recent years such as the rise of air temperature and the increase in precipitation. In particular, the rising temperature accelerates melting of glaciers and perennial snow cover and triggers permafrost degradation, and leads to the expansion of most lakes across the plateau. In addition, different distributions and types of permafrost may cause different lake variations in the southern Tibetan Plateau.

Patterns and potential drivers of dramatic changes in Tibetan lakes, 1972-2010.

Most glaciers in the Himalayas and the Tibetan Plateau are retreating, and glacier melt has been emphasized as the dominant driver for recent lake expansions on the Tibetan Plateau. By investigating detailed changes in lake extents and levels across the Tibetan Plateau from Landsat/ICESat data, we found a pattern of dramatic lake changes from 1970 to 2010 (especially after 2000) with a southwest-northeast transition from shrinking, to stable, to rapidly expanding. This pattern is in distinct contrast to the spatial characteristics of glacier retreat, suggesting limited influence of glacier melt on lake dynamics. The plateau-wide pattern of lake change is related to precipitation variation and consistent with the pattern of permafrost degradation induced by rising temperature. More than 79% of lakes we observed on the central-northern plateau (with continuous permafrost) are rapidly expanding, even without glacial contributions, while lakes fed by retreating glaciers in southern regions (with isolated permafrost) are relatively stable or shrinking. Our study shows the limited role of glacier melt and highlights the potentially important contribution of permafrost degradation in predicting future water availability in this region, where understanding these processes is of critical importance to drinking water, agriculture, and hydropower supply of densely populated areas in South and East Asia.

Application of the inundation area-lake level rating curves constructed from the SRTM DEM to retrieving lake levels from satellite measured inundation areas

Remote sensing technology has great potential for measuring lake inundation areas and lake levels, and providing important lake water quantity and quality information which can be used for improving our understanding of climate change impacts on the global water cycle, and assessing the influence of the projected future climate change on the global water resources. One remote sensing approach is to estimate lake level from satellite measured inundation area based on the inundation area-lake level rating (IALLR) curves. However, this approach is not easy to implement because of a lack of data for constructing the IALLR curves. In this study, an innovative and robust approach to construct the IALLR curves from the digital elevation model (DEM) data collected during the Shuttle Radar Topography Mission (SRTM) was developed and tested. It was shown that the IALLR curves derived from the SRTM DEM data could be used to retrieve lake level from satellite measured inundation area. Applying the constructed IALLR curve to the estimated inundation areas from 16 Landsat Thematic Mapper (TM) images, 16 lake levels of Lake Champlain in Vermont were obtained. The root mean square error (RMSE) of the estimated lake levels compared to the observed water levels at the U.S. Geological Survey (USGS) gauging station (04294500) at Burlington, Vermont is about 0.12m.

Analysis of temporal radar backscatter of rice: A comparison of SAR observations with modeling results

In this study an established microwave backscatter model is used to predict the radar backscatter behavior of rice and to understand the interaction between backscatter and the rice canopy during its growth cycle. The emphasis of this study is to understand the effect of physical plant parameters on the backscatter signatures as a function of polarization and how these signatures vary during a complete growth cycle of rice. Inputs to the backscatter model included the physical parameters of rice as obtained through field measurements. These measurements were acquired within a few days of multiple RADARSAT acquisitions of the Zhaoqing test site in southern China. RADARSAT observations and the modeling results were then compared and analyzed. The results show that the interaction mechanisms change during the rice growth cycle and polarimetric and (or) multi-polarization measurements at C band will contribute to rice monitoring programs.

Object oriented method for detection of inundation extent using multi-polarized synthetic aperture radar image

With the all-weather and day/night imaging capability, synthetic aperture radar (SAR) plays an important role in inundation extent detection. The inundation area detection using SAR will be easy as a result of the dark image tones yielded by specular reflection to the radar wave. Object oriented method (OOM) was applied to detect inundation extent using multi-polarized ENVISAT ASAR data. The traditional pixel-based methods used in information extraction and classification focus on the single pixel, so when they are applied in SAR imagery no perfect results can be achieved because of the speckle of SAR imagery. On the other hand, the pixel-based methods have limitations for detecting inundation extent and flood monitoring because of the neglect of the information of the adjacent pixels. The OOM, which no longer looks at individual pixels, but rather homogeneity areas (image objects), would be much more effective. In this paper, the OOM is applied in the ENVISAT ASAR alternative polarized (VV/VH) images using the software eCognition. The study site is located in Poyang Lake wetland, which has different inundation extent at different time. The images were segmented firstly, then the standard nearest neighbor classifier and the membership function classifier were used to classify the image objects, finally the different inundation areas were detected. The classification accuracies for two classifiers from the OOM are 95.78% and 92.24%, which are higher than that of the maximum likelihood classifier, 86.02%.

Monitoring lake-level changes in the Qinghai–Tibetan Plateau using radar altimeter data (2002–2012)

Abstract Lake-level change can be an important indicator for the water balance in the Qinghai–Tibetan Plateau (QTP). However, it is not feasible to perform continuous in-situ measurements for a large number of lakes because of the remoteness and harsh climatic conditions on this plateau. Satellite altimetry has been successfully used for monitoring water-level changes of inland lakes in recent years. In this study, water-level changes between 2002 and 2012 of 51 lakes on the QTP were monitored using multialtimeter data from Envisat/RA-2, Cryosat-2/Siral, Jason-1/Poseidon-2, and Jason-2/Poseidon-3. The water levels of 42 of the lakes have a mean rising trend of 0.275     m / a , whereas the water levels of nine lakes have a mean decreasing trend of − 0.144     m / a . Overall, the water level of these lakes had a mean increasing trend of 0.201       m / a in the past 10 years. For the lakes distributed over the entire plateau, it was found that the lake levels in different basins had different change characteristics: the lake levels in the southern plateau show a decreasing trend, whereas lake levels in the northern plateau show an increasing trend. In the central plateau, the water levels of most lakes also show an increasing trend but the water levels of a small number of the lakes show a decreasing trend. In addition, the winter or summer linear trends for the levels of individual lakes appear to be opposite to the trends observed for the yearly average. The combined use of data from several different altimeters makes the temporal resolution of lake-level change measurements higher than for those results derived using only one kind of altimeter data. However, the number of lakes monitored in this study is affected by the footprints of satellites orbits on the Earth’s surface. After taking into account anomalous lake levels, systematic elevation differences, and periodic changes in the water level, the biggest root-mean-square-error for the lakes monitored in this study is < 40     cm / a showing that the results have a high degree of accuracy. Lake-level change is mainly related to rising temperatures, increasing precipitation, and decreasing evaporation. In particular, rising temperatures accelerate the melting of glaciers and perennial snow cover and trigger permafrost degradation, leading to an increase in the water level of most lakes across the plateau.

Biomass estimation of wetland vegetation in Poyang Lake area using ENVISAT advanced synthetic aperture radar data

Abstract Biomass estimation of wetlands plays a role in understanding dynamic changes of the wetland ecosystem. Poyang Lake is the largest freshwater lake in China, with an area of about 3000     km 2 . The lake’s wetland ecosystem has a significant impact on leveraging China’s environmental change. Synthetic aperture radar (SAR) data are a good choice for biomass estimation during rainy and dry seasons in this region. In this paper, we discuss the neural network algorithms (NNAs) to retrieve wetland biomass using the alternating-polarization ENVISAT advanced synthetic aperture radar (ASAR) data. Two field measurements were carried out coinciding with the satellite overpasses through the hydrological cycle in April to November. A radiative transfer model of forest canopy, the Michigan Microwave Canopy Scattering (MIMICS) model, was modified to fit to herbaceous wetland ecosystems. With both ASAR and MIMICS simulations as input data, the NNA-estimated biomass was validated with ground-measured data. This study indicates the capability of NNA combined with a modified MIMICS model to retrieve wetland biomass from SAR imagery. Finally, the overall biomass of Poyang Lake wetland vegetation has been estimated. It reached a level of 1.09 × 10 9 , 1.86 × 10 8 , and 9.87 × 10 8     kg in April, July, and November 2007, respectively.

Poyang Lake wetland vegetation biomass inversion using polarimetric RADARSAT-2 synthetic aperture radar data

Abstract. Poyang Lake is the largest freshwater lake in China and one of the most important wetlands in the world. Vegetation, an important component of wetland ecosystems, is one of the main sources of the carbon in the atmosphere. Biomass can quantify the contribution of wetland vegetation to carbon sinks and carbon sources. Synthetic aperture radar (SAR), which can operate in all day and weather conditions and penetrate vegetation to some extent, can be used to retrieve information about vegetation structure and the aboveground biomass. In this study, RADARSAT-2 polarimetric SAR data were used to retrieve aboveground vegetation biomass in the Poyang Lake wetland. Based on the canopy backscatter model, the vegetation backscatter characteristics in the C-band were studied, and a good relation between simulated backscatter and backscatter in the RADARSAT-2 imagery was achieved. Using the backscatter model, pairs of training data were built and used to train the back propagation artificial neural network. The biomass was retrieved using this ANN and compared with the field survey results. The root-mean-square error in the biomass estimation was 45.57  g/m2. This shows that the combination of the model and polarimetric decomposition components can efficiently improve the inversion precision.

DEM generation in the densely vegetated area of Hotan, north-west China using SIR-C repeat pass polarimetric SAR interferometry

The potential use of the coherence optimization method of polarimetric SAR interferometry (Pol-InSAR) to increase the accuracy of digital elevation models (DEMs) was investigated in a densely vegetated area of the Hotan region, Xinjiang Province, north-west China, using SIR-C data. Under the same experimental conditions, the results indicate that the accuracy of a DEM generated from L-band fully polarimetric single baseline interferometric data is significantly greater than that derived from L-band HH-HH single baseline interferometric data, and the rms. error is less than 8 m.

Improved alpine grassland mapping in the Tibetan Plateau with MODIS time series: a phenology perspective

The Tibetan Plateau is primarily composed of alpine grasslands. Spatial distributions of alpine grasses, however, are not well documented in this remote, highly uninhabited region. Taking advantage of the frequently observed moderate resolution imaging spectroradiometer (MODIS) images (500-m, 8-day) in 2010, this study extracted the phenological metrics of alpine grasses from the normalized difference vegetation index time series. With the Support Vector Machine, a multistep classification approach was developed to delineate alpine meadows, steppes, and desert grasses. The lakes, permanent snow, and barren/desert lands were also classified with a MODIS scene acquired in the peak growing season. With ground data collected in the field and aerial experiments in 2011, the overall accuracy reached 93% when alpine desert grasses and barren lands were not examined. In comparison with the recently published national vegetation map, the alpine grassland map in this study revealed smoother transition between alpine meadows and steppes, less alpine meadows in the southwest, and more barren/deserts in the high-cold Kunlun Mountain in the northeast. These variations better reflected climate control (e.g. precipitation) of different climatic divisions on alpine grasslands. The improved alpine grassland map could provide important base information about this cold region under the pressure of rapidly changing climate.