Effects of climate change on vegetation dynamics of the Qinghai-Tibet Plateau, a causality analysis using empirical dynamic modeling.

Given the vital role of the Qinghai-Tibet Plateau (QTP) as water tower in Asia and regulator for regional and even global climate, the relationship between climate change and vegetation dynamics on it has received considerable focused attention. Climate change may influence the vegetation growth on the plateau, but clear empirical evidence of such causal linkages is sparse. Herein, using datasets CRU-TS v4.04 and AVHHR NDVI from 1981 to 2019, we quantify causal effects of climate factors on vegetation dynamics with an empirical dynamical model (EDM) -- a nonlinear dynamical systems analysis approach based on state-space reconstruction rather than correlation. Results showed the following: (1) climate change promotes the growth of vegetation on the QTP, and specifically, this favorable influence of temperature is stronger than precipitation's; (2) the direction and strength of climate effects on vegetation varied over time, and the effects are seasonally different; (3) a significant increase in temperature and a slight increase in precipitation are beneficial to vegetation growth, specifically, NDVI will increase within 2% in the next 40 years with the climate trend of warming and humidity. Besides the above results, another interesting finding is that the two seasons in which precipitation strongly influence vegetation in the Three-River Source region (part of the QTP) are spring and winter. This study provides insights into the mechanisms by which climate change affects vegetation growth on the QTP, aiding in the modeling of vegetation dynamics in future scenarios.

A novel autoencoder approach to feature extraction with linear separability for high-dimensional data.

Feature extraction often needs to rely on sufficient information of the input data, however, the distribution of the data upon a high-dimensional space is too sparse to provide sufficient information for feature extraction. Furthermore, high dimensionality of the data also creates trouble for the searching of those features scattered in subspaces. As such, it is a tricky task for feature extraction from the data upon a high-dimensional space. To address this issue, this article proposes a novel autoencoder method using Mahalanobis distance metric of rescaling transformation. The key idea of the method is that by implementing Mahalanobis distance metric of rescaling transformation, the difference between the reconstructed distribution and the original distribution can be reduced, so as to improve the ability of feature extraction to the autoencoder. Results show that the proposed approach wins the state-of-the-art methods in terms of both the accuracy of feature extraction and the linear separabilities of the extracted features. We indicate that distance metric-based methods are more suitable for extracting those features with linear separabilities from high-dimensional data than feature selection-based methods. In a high-dimensional space, evaluating feature similarity is relatively easier than evaluating feature importance, so that distance metric methods by evaluating feature similarity gain advantages over feature selection methods by assessing feature importance for feature extraction, while evaluating feature importance is more computationally efficient than evaluating feature similarity.