ABSTRACT
Carbon use efficiency (CUE) is a key element in the vegetation carbon cycle, and determines how vegetation allocates carbon. Here, our research provides the spatio-temporal variations of CUE on the Tibetan Plateau (TP) based on ensemble simulations from 12 terrestrial ecosystem models. Moreover, the experimental design of simulations adds one time-varying driver at a time, thus quantitative analysis of the response of CUE to climate factors (i.e., temperature, precipitation and radiation), land use and land cover change (LULCC), and CO2 fertilization can be investigated. Results show that average CUE value of the multi-model simulations (0.583 ± 0.064) on the TP is slightly lower than that derived from the satellite-based product, the Moderate Resolution Imaging Spectroradiometer (0.646). However, CUE varies greatly among models due to differences in simulating plant photosynthetic productivity and respiratory rate, with range of 0.489-0.661. LULCC and CO2 fertilization contribute 4.24% and 0.79% of the annual mean CUE, respectively. Among the climatic factors, temperature and precipitation have positive correlations with CUE over most areas of the TP while solar radiation shows a negative impact.
ABSTRACT
Excess nutrients from fertilizer application, pollution discharge, and water regulations outflow through rivers from lands to oceans, seriously impacting coastal ecosystems. A reasonable representation of these processes in land surface models and River Transport Models (RTMs) is very important for understanding human-environment interactions. In this study, the schemes of riverine dissolved inorganic nitrogen (DIN) transport and human activities including nitrogen discharge and water regulation, were synchronously incorporated into a land surface model coupled with a RTM. The effects of anthropogenic nitrogen discharge on the DIN transport in rivers were studied based on simulations of the period 1991-2010 throughout the entire world, conducted using the developed model, which had a spatial resolution of about 1° for land processes and 0.5° for river transport, and data on fertilizer application, point source pollution, and water use. Our results showed that rivers in western Europe and eastern China were seriously polluted, on average, at a rate of 5,000-15,000 tons per year. In the Yangtze River Basin, the amount of point source pollution in 2010 was about four times more than that in 1991, while the amount of fertilizer used in 2010 doubled, which resulted in the increased riverine DIN levels. Further comparisons suggested that the riverine DIN in the USA was affected primarily by nitrogen fertilizer use, the changes in DIN flow rate in European rivers was dominated by point source pollution, and rivers in China were seriously polluted by both the two pollution sources. The total anthropogenic impact on the DIN exported to the Pacific Ocean has increased from 10% to 30%, more significantly than other oceans. In general, our results indicated that incorporating the schemes of nitrogen transport and human activities into land surface models could be an effective way to monitor global river water quality and diagnose the performance of the land surface modeling.
ABSTRACT
An inverse source estimation method is proposed to reconstruct emission rates using local air concentration sampling data. It involves the nonlinear least squares-based ensemble four-dimensional variational data assimilation (NLS-4DVar) algorithm and a transfer coefficient matrix (TCM) created using FLEXPART, a Lagrangian atmospheric dispersion model. The method was tested by twin experiments and experiments with actual Cs-137 concentrations measured around the Fukushima Daiichi Nuclear Power Plant (FDNPP). Emission rates can be reconstructed sequentially with the progression of a nuclear accident, which is important in the response to a nuclear emergency. With pseudo observations generated continuously, most of the emission rates were estimated accurately, except under conditions when the wind blew off land toward the sea and at extremely slow wind speeds near the FDNPP. Because of the long duration of accidents and variability in meteorological fields, monitoring networks composed of land stations only in a local area are unable to provide enough information to support an emergency response. The errors in the estimation compared to the real observations from the FDNPP nuclear accident stemmed from a shortage of observations, lack of data control, and an inadequate atmospheric dispersion model without improvement and appropriate meteorological data. The proposed method should be developed further to meet the requirements of a nuclear emergency response.
