Votre recherche

A bstract Developing models to predict tree mortality using data from long‐term repeated measurement data sets can be difficult and challenging due to the nature of mortality as well as the effects of dependence on observations. Marginal (population‐averaged) generalized estimating equations (GEE) and random effects (subject‐specific) models offer two possible ways to overcome these effects. For this study, standard logistic, marginal logistic based on the GEE approach, and random logistic regression models were fitted and compared. In addition, four model evaluation statistics were calculated by means of K ‐fold cross‐valuation. They include the mean prediction error, the mean absolute prediction error, the variance of prediction error, and the mean square error. Results from this study suggest that the random effects model produced the smallest evaluation statistics among the three models. Although marginal logistic regression accommodated for correlations between observations, it did not provide noticeable improvements of model performance compared to the standard logistic regression model that assumed impendence. This study indicates that the random effects model was able to increase the overall accuracy of mortality modeling. Moreover, it was able to ascertain correlation derived from the hierarchal data structure as well as serial correlation generated through repeated measurements.

Abstract A method to assess firn compaction using data collected with the Airborne SAR (Synthetic Aperture Radar)/Interferometric Radar Altimeter System (ASIRAS) is developed. For this, we develop a dynamical firn-compaction model that includes meltwater retention. Based on the ASIRAS data, which show internal layers as annual horizons in the uppermost firn, the method relies on inferring the age/ depth (internal layers) information from the radar data using a Monte Carlo inversion technique to tune in parallel both the firn model and the atmospheric forcing parameters (temperature and accumulation). The model is validated against two firn cores, and it is shown that applying both firn densities and age/ depth information for the inversion gives the most accurate understanding of model biases. The method is then applied to a 67 km section of the EGIG line forced by atmospheric output from a regional climate model using only age/depth information in the inversion step. The layers traced by the ASIRAS data are modeled with a root-mean-square error of 9 cm, which is within the estimated error of the layer tracing. This gives us confidence in applying observed annual layering from firn radar data to assess firn compaction; however, the study also indicates that our firn-model-tuning parameters are site-dependent and cannot be parameterized by temperature and accumulation alone.

Alongside global warming, droughts are expected to increase in frequency, severity, and extent in the near future, which will likely result in significant impacts on forest growth, production, structure, composition, and ecosystem services. However, due to spatial and temporal characteristics, it is difficult to monitor and assess the potential effects of droughts. Remote sensing can provide an effective way to obtain real-time conditions of forests affected by drought and offer a range of spatial and temporal insights into drought-induced changes to forest ecosystem structure, function, and services. Remote sensing is rapidly developing as more satellites are launched. In situ and remotely sensed data fusion techniques have achieved notable success in assessing drought-induced damage to forests and carbon cycles. Even so, constraints still exist when using satellite data. The objectives of this review are to (1) briefly review existing data sources and methods of remote sensing; (2) synthesize current applications and contributions of remote sensing in monitoring and estimating impacts of droughts on forest ecosystems; and (3) highlight research gaps and future challenges.

Plants interact to the seasonality of their environments, and changes in plant phenology have long been regarded as sensitive indicators of climatic change. Plant phenology modeling has been shown to be the simplest and most useful tool to assess phenol–climate shifts. Temperature, solar radiation, and water availability are assumed to be the key factors that control plant phenology. Statistical, mechanistic, and theoretical approaches have often been used for the parameterization of plant phenology models. The statistical approaches correlate the timing of phenological events to environmental factors or heat unit accumulations. The approaches have the simplified calculation procedures, correct phenological mechanism assumptions, but limited applications and predictive abilities. The mechanistic approaches describe plant phenology with the known or assumed “cause–effect relationships” between biological processes and key driving variables. The mechanistic approaches have the improved parameter processes, realistic assumptions, broad applications, and effective predictions. The theoretical approaches assume cost–benefit tradeoff strategies in trees. These methods are capable of capturing and quantifying the potential impacts and consequences of global climate change and human activity. However, certain limitations still exist related to our understanding of phenological mechanisms in relation to (1) interactions between plants and their specific climates, (2) the integration of both field observational and remote sensing data with plant phenology models across taxa and ecosystem type, (3) amplitude clarification of scale-related sensitivity to global climate change, and (4) improvements in parameterization processes and the overall reduction of modeling uncertainties to forecast impacts of future climate change on plant phenological dynamics. To improve our capacity in the prediction of the amplitude of plant phenological responses with regard to both structural and functional sensitivity to future global climate change, it is important to refine modeling methodologies by applying long-term and large-scale observational data. It is equally important to consider other less used but critical factors (such as heredity, pests, and anthropogenic drivers), apply advanced model parameterization and data assimilation techniques, incorporate process-based plant phenology models as a dynamic component into global vegetation dynamic models, and test plant phenology models against long-term ground observations and high-resolution satellite data across different spatial and temporal scales.

Abstract Recent studies have suggested that tropical forests may not be resilient against climate change in the long term, primarily owing to predicted reductions in rainfall and forest productivity, increased tree mortality, and declining forest biomass carbon sinks. These changes will be caused by drought‐induced water stress and ecosystem disturbances. Several recent studies have reported that climate change has increased tree mortality in temperate and boreal forests, or both mortality and recruitment rates in tropical forests. However, no study has yet examined these changes in the subtropical forests that account for the majority of China's forested land. In this study, we describe how the monsoon evergreen broad‐leaved forest has responded to global warming and drought stress using 32 years of data from forest observation plots. Due to an imbalance in mortality and recruitment, and changes in diameter growth rates between larger and smaller trees and among different functional groups, the average DBH of trees and forest biomass have decreased. Sap flow measurements also showed that larger trees were more stressed than smaller trees by the warming and drying environment. As a result, the monsoon evergreen broad‐leaved forest community is undergoing a transition from a forest dominated by a cohort of fewer and larger individuals to a forest dominated by a cohort of more and smaller individuals, with a different species composition, suggesting that subtropical forests are threatened by their lack of resilience against long‐term climate change.