Hyperspectral Study of Estimating Nitrification Microorganism in Wetland Soils.

Nitrogen cycle is an important process in the circle of soil ecosystem elements, and nitrification has significant effect on soil nitrogen cycling. The main completer of nitrification is nitrification microbial communities. Soil microorganisms are vital components of wetland ecosystem. They can indicate the variations of wetland ecological environment, and this helps us to have the correct understanding of nitrogen cycle and pollution purification function in wetland ecosystem. This paper tries to study nitrification microbial communities in wetland soils from the perspective of hyperspectral remote sensing technology, based on the monitoring mechanisms of soil nitrogen spectrum. The study explores hyperspectral estimation techniques for nitrification microbial communities in wetland soils, and it can provide a new technical approach to estimate the temporal and spatial distribution of nitrification microbial communities. The study adopted most probable number method (MPN) to count the numbers of ammonia oxidizing bacteria and nitrite oxidizing bacteria respectively, which were main completers of two independent stages in nitrification. And the total results of both count measurements were used as the values of soil nitrification microorganisms for each sampling area. The estimation models of nitrification microorganism and total nitrogen in wetland soils were developed respectively using spectral transformation techniques, such as log-transformed spectra (LR), first derivative (FD), second derivative (SD), continuum removal (CR) and band depth (BD), and modeling methods, such as stepwise multiple linear regression (SMLR) and partial least-squares regression (PLSR) based on the bootstrap technology. The results indicated that the selected estimation bands of nitrification microorganism and total nitrogen were close (especially for original spectral data (R) and SD spectra) when the modeling method of bootstrap SMLR was used. Compared to the bootstrap SMLR, the bootstrap PLSR achieved higher accuracies for estimating nitrification microorganism and total nitrogen in wetland soils. The spectral transformation technique of SD combined with the modeling method of bootstrap PLSR yielded the highest estimation accuracy to predict nitrification microorganism in wetland soils. The CR spectral data combined with bootstrap PLSR produced the highest estimation accuracy to predict total nitrogen content in wetland soils.

[Progress in inversion of vegetation nitrogen concentration by hyperspectral remote sensing].

Nitrogen is the necessary element in life activity of vegetation, which takes important function in biosynthesis of protein, nucleic acid, chlorophyll, and enzyme etc, and plays a key role in vegetation photosynthesis. The technology about inversion of vegetation nitrogen concentration by hyperspectral remote sensing has been the research hotspot since the 70s of last century. With the development of hyperspectral remote sensing technology in recent years, the advantage of spectral bands subdivision in a certain spectral region provides the powerful technology measure for correlative spectral characteristic research on vegetation nitrogen. In the present paper, combined with the newest research production about monitoring vegetation nitrogen concentration by hyperspectral remote sensing published in main geography science literature in recent several years, the principle and correlated problem about monitoring vegetation nitrogen concentration by hyperspectral remote sensing were introduced. From four aspects including vegetation nitrogen spectral index, vegetation nitrogen content inversion based on chlorophyll index, regression model, and eliminating influence factors to inversion of vegetation nitrogen concentration, main technology methods about inversion of vegetation nitrogen concentration by hyperspectral remote sensing were detailedly introduced. Correlative research conclusions were summarized and analyzed, and research development trend was discussed.

[Monitoring gas concentration from carbon emissions by remote sensing].

Global climate warming has become the focus question of international global climate change research, and is an important factor influencing world economy, political situation, and ecological environment. Produced carbon emission gases such as CO2, CH4, N2O, etc. caused by human activity are the main reason for global warming. In order to forecast future climate change and construct accurate carbon cycle model, monitoring accuracy of gas concentration from carbon emission must be improved. In the present paper, the newest progress in the international research results about monitoring gas concentration from carbon emissions by remote sensing was considered, monitoring method for carbon emissions was introduced, and remotely sensed monitoring technology about gas concentration from carbon emissions (including thermal infrared, sun spectrum, active remote sensing monitoring technology) was stated. In detail, several present and future satellite sensors were introduced (including TOVS, AIRS, IASI, SCIAMACHY, GOSAT, OCO, A-SCOPE and ASCENDS), and monitoring results achieved by these sensors were analyzed.

[Progress in research on land cover products of MODIS].

Remote sensing technology has been rapidly developed in recent decades, and has been widely used in ecology and environment field. MODIS is a new data source, and in its many products, land cover product is an important product, and it has often been used in global and regional models. In the present review, the procedure of producing land cover product is clearly discussed, and the feature of the classification based network or decision tree is introduced. The paper emphasized the importance of direction information in classification, detailed introduction of change vector analysis methods and land cover change detection based artificial nerve network, analyzed global 17 land cover types defined by IGBP, and compared with other 3 classification systems.

[Analysis of vegetation dynamics over Liaoning province based on remote sensing data].

In the present paper, the authors used normalized difference vegetation index (NDVI) data derived from NOAA AVHRR sensor to analyze spatial heterogeneity and temporal dynamics of Liaoning province during the past two decades. A set of 2 292 spatially distributed NDVI values were analyzed to investigate obvious deviations by the mean-monthly values from 1982 to 2001. Various statistical analyses including minimum, mean and maximum values, coefficient of variation (CV), standardized anomalies (Z-scores), and 36-month running mean were used for monthly NDVI values to research spatial and temporal variations in vegetation. In Liaoning province, the authors found the strong seasonal oscillations during plants growing period, the maximum value of NDVI appeared in July-August, and seasonal variation ranged from 6% to 14% of CV value. Vegetation greenness kept upward trend from 1984 to 1990, but showed downward trend from 1991 to 1998. Vegetation greenness followed an interannual oscillation period of 7-8 years. The authors also found that the variation of NDVI peak along latitude direction was 20%-25% greater in 1991-1999 than in 1982-1990 in dry season across Liaoning province. The conclusions of this paper suggest that the patterns of vegetation variability in Liaoning province were caused by enhanced aridity occurring over the last decade of the 20th century.

[Analysis of vegetation spatial and temporal variations in Qinghai Province based on remote sensing].

1 km MODIS NDVI time series data combining with decision tree classification, supervised classification and unsupervised classification was used to classify land cover type of Qinghai Province into 14 classes. In our classification system, sparse grassland and sparse shrub were emphasized, and their spatial distribution locations were labeled. From digital elevation model (DEM) of Qinghai Province, five elevation belts were achieved, and we utilized geographic information system (GIS) software to analyze vegetation cover variation on different elevation belts. Our research result shows that vegetation cover in Qinghai Province has been improved in recent five years. Vegetation cover area increases from 370047 km2 in 2001 to 374576 km2 in 2006, and vegetation cover rate increases by 0.63%. Among five grade elevation belts, vegetation cover ratio of high mountain belt is the highest (67.92%). The area of middle density grassland in high mountain belt is the largest, of which area is 94 003 km2. Increased area of dense grassland in high mountain belt is the greatest (1280 km2). During five years, the biggest variation is the conversion from sparse grassland to middle density grassland in high mountain belt, of which area is 15931 km2.