Rice is one of the most important staple food sources worldwide. Effective and cheap monitoring of rice planting areas is demanded by many developing countries. This study proposed a weakly supervised paddy rice mapping approach based on long short-term memory (LSTM) network and dynamic time warping (DTW) distance. First, standard temporal synthetic aperture radar (SAR) backscatter profiles for each land cover type were constructed on the basis of a small number of field samples. Weak samples were then labeled on the basis of their DTW distances to the standard temporal profiles. A time series feature set was then created that combined multi-spectral Sentinel-2 bands and Sentinel-1 SAR vertical received (VV) band. With different combinations of training and testing datasets, we trained a specifically designed LSTM classifier and validated the performance of weakly supervised learning. Experiments showed that weakly supervised learning outperformed supervised learning in paddy rice identification when field samples were insufficient. With only 10% of field samples, weakly supervised learning achieved better results in producer’s accuracy (0.981 to 0.904) and user’s accuracy (0.961 to 0.917) for paddy rice. Training with 50% of field samples also presented improvement with weakly supervised learning, although not as prominent. Finally, a paddy rice map was generated with the weakly supervised approach trained on field samples and DTW-labeled samples. The proposed data labeling approach based on DTW distance can reduce field sampling cost since it requires fewer field samples. Meanwhile, validation results indicated that the proposed LSTM classifier is suitable for paddy rice mapping where variance exists in planting and harvesting schedules.

Alpine treeline, as a prominent ecological boundary between forested mountain slopes and alpine meadow/shrub, is highly complex in altitudinal distribution and sensitive to warming climate. Great efforts have been made to explore their distribution patterns and ecological mechanisms that determine these patterns for more than 100 years, and quite a number of geographical and ecophysiological models have been developed to correlate treeline altitude with latitude or a latitude related temperature. However, on a global scale, all of these models have great difficulties to accurately predict treeline elevation due to the extreme diversity of treeline site conditions. One of the major reasons is that “mass elevation effect” (MEE) has not been quantified globally and related with global treeline elevations although it has been observed and its effect on treeline elevations in the Eurasian continent and Northern Hemisphere recognized. In this study, we collected and compiled a total of 594 treeline sites all over the world from literatures, and explored how MEE affects global treeline elevation by developing a ternary linear regression model with intra-mountain base elevation (IMBE, as a proxy of MEE), latitude and continentality as independent variables. The results indicated that IMBE, latitude and continentality together could explain 92% of global treeline elevation variability, and that IMBE contributes the most (52.2%), latitude the second (40%) and continentality the least (7.8%) to the altitudinal distribution of global treelines. In the Northern Hemisphere, the three factors’ contributions amount to 50.4%, 45.9% and 3.7% respectively; in the south hemisphere, their contributions are 38.3%, 53%, and 8.7%, respectively. This indicates that MEE, virtually the heating effect of macro-landforms, is actually the most significant factor for the altitudinal distribution of treelines across the globe, and that latitude is relatively more significant for treeline elevation in the Southern Hemisphere probably due to fewer macro-landforms there.

The Rocky Mountains are the highest and most extensive in North America. Their mass-elevation effect (MEE) should be strong and responsible for the high elevations of treelines in the interior of the mountain range. However, the MEE has been little known. We calculated the temperature difference (ΔT) between the inner and outer Rocky Mountains at the same elevation and defined it as the magnitude of MEE. Results show a mean ΔT of 1.8°C, with the highest ΔT occurring in Colorado and the basins of southern Wyoming. The MEE can be modelled with three factors: mountain base elevation (MBE), latitude, and hygric continentality. The model has a high explanatory power of 68.90%, and the three factors contribute 45.65, 36.05, and 18.03%, respectively. MBE contributed most to MEE, both in the whole range (45.65%) and in the Colorado Rocky Mountains (55.21%). Treeline is higher in the interior than the outside of the mountain range, by 700–1400 m. This difference coincides well with air temperature differences. This study developed a quantitative model for the MEE of the Rocky Mountains and improves our understanding of the intra-mountain ecology and high treelines of the Rocky Mountains.