Grassland in northern Tibet plays an important role in the eco-security of the Qinghai-Tibet Plateau and the restoration of deserte@and degraded grassland is now a focus for governments. We used remote sensing, simulations and field surveys to analyze the current status, trends and causes of grassland degradation across northern Tibet. We develop several recovery models for degraded grassland based on field experiments in the region. We found that slightly degraded grassland covers 62% and that moderate to severely degraded grassland occupied 15.1% in the Chang Tang Plateau. The amount of degraded alpine steppe increased from 1991, and the amount of area classified as severely degraded increased sharply from 2000. The cause of degraded steppe in northwestern Tibet may be the result of warming and an arid climate; the cause of severe degradation in mid and eastern regions was mainly from overgrazing. Three restoration models are proposed for different levels of degradation： ＂enclosures＂ for slightly degraded areas, ＂enclosures with fertilization＂ for moderately degraded areas, and ＂enclosure with oversowing and fertilization＂ for severely degraded areas.
Drought may impact the net ecosystem exchange of CO2 （NEE） between grassland ecosystems and the atmosphere during growth seasons. Here, carbon dioxide exchange and controlling factors in alpine grassland under drought stress in the hinterland of Tibetan Plateau （Damxung, Tibet, China） were investigated. Data were obtained using the covariance eddy technique in 2009. Severe drought stress appeared in the early growing season （May to early July） and September. Drought conditions during the early growing season limited grass production and the green leaf area index （GLAD increased slowly, with an obvious decline in June. When encountering severe water stress, diurnal patterns of NEE in the growth season altered with a peak carbon release around 16：00 h or a second carbon uptake period before sunset. NEE variations in daytime related most closely with O other than PAR when daily averaged @〈0.1 m3 m 3. Seasonal patterns of gross primary production （GPP） and NEE were also influenced by drought： the maximum and minimum of daily-integrated NEE were 0.9 g C m-2 d-1 on 3 July 2009, and -1.3 g C m-2 d-1 on 12 August 2009 with a GPP peak （-2.3 g C m-2 d-1） on the same day, respectively. Monthly NEE from May to July remained as carbon release and increased gradually; peak values of monthly NEE and GPP both appeared in August, but that of ecosystem respiration （R^co） was reached in July. Annual NEE, GPP and Reco of the alpine grassland ecosystem were 52.4, -158.1 and 210.5 g C m-2, respectively. Therefore, the grassland was a moderate source of COs to the atmosphere in this dry year. Interannual variation in NEE was likely related to different water conditions in the growing season. The three greatest contributors to seasonal variation in NEE, GPP and R^co respectively were GLAI〉Ta〉O, GLAI〉O〉PPT, and Ta〉GLAI〉PAR. Seasonality of GLAI explained 60.7% and 76.1% of seasonal variation in NEE and GPP, respectively. GPP or NEE was more sensitive than Reco to variation in GLAI, and ecosystem
