Vegetation is known to be sensitive to both climate change and anthropogenic disturbance in the karst region. However, the relationship between an abrupt change in vegetation and its driving factors is unclear at multiple timescales. Based on the non-parametric Mann-Kendall test and the ensemble empirical mode decomposition (EEMD) method, the abrupt changes in vegetation and its possible relationships with the driving factors in the karst region of southwest China during 1982–2015 are revealed at multiple timescales. The results showed that: (1) the Normalized Difference Vegetation Index (NDVI) showed an overall increasing trend and had an abrupt change in 2001. After the abrupt change, the greening trend of the NDVI in the east and the browning trend in the west, both changed from insignificant to significant. (2) After the abrupt change, at the 2.5-year time scale, the correlation between the NDVI and temperature changed from insignificantly negative to significantly negative in the west. Over the long-term trend, it changed from significantly negative to significantly positive in the east, but changed from significantly positive to significantly negative in the west. The abrupt change primarily occurred on the long-term trend. (3) After the abrupt change, 1143.32 km² farmland was converted to forests in the east, and the forest area had significantly increased. (4) At the 2.5-year time scale, the abrupt change in the relationships between the NDVI and climate factors was primarily driven by climate change in the west, especially rising temperatures. Over the long-term trend, it was caused by ecological protection projects in the east, but by rising temperatures in the west. The integration of the abrupt change analysis and multiple timescale analysis help assess the relationship of vegetation changes with climate changes and human activities accurately and comprehensively, and deepen our understanding of the driving mechanism of vegetation changes, which will further provide scientific references for the protection of fragile ecosystems in the karst region.

Spartina alterniflora has become the main invasive plant along the Chinese coast and now threatens the local ecological environment. Accurately monitoring the distribution of S. alterniflora is urgent and essential for developing cost-effective control strategies. In this study, we applied the One-Class Classification (OCC) methods of Maximum entropy (Maxent) and Biased Support Vector Machine (BSVM) based on Landsat time-series imagery to detect the species on the middle coast of Jiangsu in east China. We conducted four experimental setups (i.e., single-scene analysis, time-series analysis, Normalized Difference Vegetation Index (NDVI) time-series analysis and a compressed time-series analysis), using OCC methods to recognize the species. Then, we tested the performance of a compressed time-series model for S. alterniflora detection and evaluated the expansibility of this approach when it was applied to a larger region. Our principal findings are as follows: (1) Maxent and BSVM performed equally well, and Maxent appeared to have a more balanced performance over the summer months; (2) the Maxent model with the Default Parameter Set (Maxent-DPS) showed a slightly higher accuracy and more overfitting than Maxent with the Akaike Information Criterion corrected for small samples sizes (AICc)-selected parameter set model, but a t-test found no significant difference between these two settings; (3) April and December were deemed to be important periods for the detection of S. alterniflora; (4) a compressed time-series analysis model—including only three variables (December NDVI, March green and the third Principal Component in January, PC3)—yielded higher accuracy than single-scene analyses, which indicated that time-series analysis can better detect S. alterniflora than single-scene analyses; and (5) the Maxent model using the reconstructed optimal variables and 70 training samples over a larger region produced encouraging results with an overall accuracy of 90.88% and a Kappa of 0.78. The one-class classification method combined with a phenology-based detection strategy is therefore promising for the application of the long-term detection of S. alterniflora over extended areas.

The trend of vegetation phenology dynamics is of crucial importance for understanding vegetation growth and its responses to climate change. However, it remains unclear how the trends of vegetation phenology changed over the past decades. By analyzing phenology data including start (SOS), end (EOS) and length (LOS) of growth season with the Ensemble empirical mode decomposition (EEMD), we revealed the trend evolution of vegetation phenology in China during 1981-2016. Our study suggests that: (1) On the national scale, with EEMD method, the change rates of SOS and LOS were increasing with time, while that of EOS was decreasing. Moreover, the EEMD rates of SOS and LOS exceeded the linear rates in the early-2000s, while that of EOS dropped below the linear rate in the mid-1980s. (2) For each phenological event, the shifted trends took up a large area (~30%), which was close to the sum of all monotonic trends, but more than any monotonic trend type. The shifted trends mainly occurred in the north-eastern China, eastern Qinghai-Tibetan Plateau, eastern Sichuan Basin, North China Plain and Loess Plateau. (3) For each phenological event, the areas in the high-latitude experienced the contrary trends to the other. The amplitude and frequencies of phenology variation in the mid-latitude were stronger than those in the high-latitude and low-latitude. Meanwhile, LOS in the high-latitude was induced by SOS. (4) For each phenological event, the trend evolution varying with longitudes can be divided into eastern region (east of 121°E), central region (92°E–121°E) and western region (west of 92°E) based on the evolution of trends varying with longitudes. The east experienced a delayed SOS and a shorten LOS, which was different from the other areas. The magnitude of delayed trends in EOS and the prolonged trends in LOS were stronger from east to west as longitudes changes. The variation characteristics of LOS with longitude were mainly caused by SOS in the eastern region and by SOS and EOS together in the western and central region. (5) Each land cover types, except Needleleaf Forests, experienced the same trends. For most land cover types, the advance of SOS, delay of EOS and extension of LOS began in the 1980s, the 1990s, and the 1990s, respectively and enhanced several times. Moreover, the magnitudes of Grasslands in SOS and Evergreen Broadleaf Forest in EOS were much greater than the others, while that of croplands was the smallest in each phenological event. Our results showed that the analysis of trend evolution with nonlinear method is very important to accurately reveal the variation characteristics of phenology trends and to extract the inherent trend shifts.