After reviewing some current common ecosystem service classification systems, it is found that current classification systems were not able to reflect the fact of different ecosystem types often produce different services due to their unique components, structures and environments. The authors used ecosystem components to create a new ecosystem service classification system based on transmission medium. Ecosystem services are classified into water-transmitted, rock and soil-transmitted, air-transmitted, organism-transmitted and whole ecosystemtransmitted services. Some water-transmitted services include water yield, water purifying, water regulation etc.; rock and soil-transmitted services include soil retention, soil formation, geologic hazard prevention etc.; airtransmitted services include carbon sequestration, air purifying, climate regulation etc.; organism-transmitted services include food production, pollination, pest bio-control etc.; last but not least, whole ecosystem-transmitted services include aesthetic information, recreation, education etc. The proposed classification system may help differentiating services provided by different ecosystem types and identifying some common services provided by the same ecosystem type.

For the land use demands of Jing-Jin-Ji urban agglomeration cooperative development strategies, four kinds of land use policies, i.e. status quo continuation, food security, nature protection and urban expansion were made. A modified Cobb-Douglas utility function was developed to quantify the influence of different policies on various land use demands, and a CLUE-S model was built to simulate the spatial and temporal evolution of Jing-Jin-Ji land use under different policies. Results show that, compared with year 2010, great changes occur on land use areas and patterns in the year 2020 under different policies. “Status quo continuation” is characterized by the continuous expansion of urban groups, increasing the construction land by 2280 km². “Food Security” significantly increases the arable land by 3611.4 km², while reducing the ecological land including forest and grassland. “Nature protection” greatly reduces the area of arable land by 3082.13 km², while increasing forest, grassland and water area by 3726.4 km². “Urban expansion” substantially increases the construction land by 3375 km², while decreasing other types of land use. Spatially, every land use policy has its significant regional characteristics in land use conversion. The increase of construction land always comes together with the decrease of arable land, which tends to occur in the existing urban surroundings. The increase of ecological land is more often seen in Bashang plateau, Yanshan Mountains as well as Taihang Mountains. This study has great reference value in designing optimal land use policies, especially in the gradual implementation stage of the Jing-Jin-Ji collaborative development strategies.

Based on the GWR (geographically weighted regression) model supported by ArcGIS, the research explores the multi-scale relationship between vegetation change, climatic factors, and the sensitivity of vegetation to climate factors using AVHRR vegetation cover data combined with temperature and precipitation data in China from 1982–2010. Compared with the general linear regression (Ordinary Least Square, OLS) model, GWR gives a much better fitting result, with the goodness of fit increased from 0.3 to 0.6. The relationship between NDVI, annual rainfall, and average annual temperature has a significant spatial heterogeneity. Regression coefficients of climatic factors decrease from north to south and are higher in the northwest dry region of China. Temperature is more influential than rainfall on NDVI in most areas of China. Each ecological zone has different spatial scales when NDVI and the climatic factors maintain a stable relationship.

Based on Meta-analysis, the ecosystem service value regression models for cultivated land, forest land, pasture, garden land, waters and unused land were established respectively. In addition to time and value method, the geographical division and socio-economic factors (including population density and GDP per capita) were also incorporated into the models. The performance of model indicates that these factors play a significant role in explaining ecosystem service value change. Using value transfer approach, the authors valued ecosystem service of land use types in Beijing, Tianjin, and 11 cities in Hebei Province. The results show that land use types ranked by ecosystem service value are waters, forest land, pasture, garden land, cultivated land and unused land. The ecosystem service values per unit area of cultivated land, forest land, pasture, garden land and water area in Beijing and Tianjin are more than other areas. The research results on the one hand can enrich the methodology and technology of ecosystem service valuation, on the other hand provide scientific support for land use sustainable management in study area.

The scale, evaluation index, method, the application in management and decision-making are reviewed through the published achievement of cultural ecosystem services (CES), and the research tendency is predicted. Perfecting the research framework, promoting the development of participatory mapping and strengthening the interaction between services are considered the most important topics in the future.

Taking Shenzhen, Guangdong Province, China as an example, CLUE-S model was used to simulate its urban sprawl in 2020. An evaluation system was established based on indicators from speed, strength, morphology, social economic benefits and ecological benefits of urban sprawl. By using SOFM network, the authors zoned the region and evaluated the urban sprawl for each zone. Main conclusions include that 1) Shenzhen can be divided into four regions according to future (2020) urban sprawl, i.e. the highly urbanized zone, the developing zone, the potential zone and the ecology priority zone, and therefore reasonable adjustments of development strategy need to be made; 2) by 2020, the overall quality of urban sprawl in Shenzhen will be high according to current trend, however, problems such as the low efficiency of land use and the incoordination of social-economic and ecological benefits still exist; 3) based on CLUE-S and SOFM, combined with urban sprawl comprehensive evaluation index system, the future urban sprawl can be predicted and analyzed effectively.

The authors put forward the concept of vulnerability of ecosystem services, and carries a case study in Beijing-Tianjin-Hebei Area. The changes of ecosystem services were calculated from 2001 to 2009 in this region, and radial basis function network was used to assess the potential impacts of ecosystem services to land use change and the adaptive capacity of human-beings based on the 9 training data chosen from 164 counties. The results show that the average ecosystem services value of study area has declined 184.36 yuan/ha from 2001 to 2009, and the positive potential impacts of ecosystem services to future land use change declines from southeast to northwest, and the middle shows the least; the districts near big cities such as Beijing, Tianjin and Tangshan have the best adaptive capacity, which has a positive correlation with GDP; and by integrating potential impacts and adaptive capacity, the study gives the spatial pattern of the vulnerability of ecosystem services: the districts near Bohai Sea turn out to be the least vulnerable; the plain in the southeast and mountains in the northwest are worse; and the districts near Taihang Mountains in the west are the most vulnerable. Of all the residents living in this region, 92.4% range between 2 to 4 in the classification of vulnerability.

The authors present an analysis of how scale issues affect a cellular automata model of land use change developed for a research area in Longhua Town, Shenzhen City. The scale dependence of the model is explored by varying the resolution of the input data in 1990 used to calibrate the model and changing the length of model simulating time. To explore the impact of these scale relationships the model is run with input datasets constructed at the following spatial resolutions: 30, 60, 90, 120, 150, 180, 210 and 240 m for simulating land use in 1995 and 2000. Three kinds of indicator, i.e. point by point accuracy, Kappa and real change accuracy are used to assess the scale sensitivity of the model. The results show that 1) the more fine the cell sizes are, the higher the accuracy of the simulation results; 2) path dependence of the isolated cells is an important source of the spatial scale sensitivity of CA model; 3) the specific geographical process in different periods of time is an important source of the temporal sensitivity scale of CA model. The results have great significance for the scale selection of CA model.

The authors investigated the temporal and spatial patterns of the virtual water content (/kg) and total virtual water requirements for rice, wheat, maize and soybeans in Liaoning Province, Jilin Province and Heilongjiang Province in Northeastern China from 1988 to 2008. The results show that soybeans had the highest water demand per kilogram, followed by wheat, rice and maize. Spatial variations were evident. Per kilogram rice consumes more water in Heilongjiang, while wheat in Jilin, maize and soybeans in Liaoning. In Heilongjiang, rice, wheat and soybeans had the highest total virtual water needs, 8.60×10⁹, 3.98×10⁹ and 1.22×10¹⁰m³, respectively. Whereas the largest total virtual water for maize (1.08×10¹⁰m³) appeared in Jilin Province. During the studied period, water demands per kilogram of soybeans and wheat fluctuated, while for rice and maize the demands were stable. As to total virtual water demand, it increased for maize and rice, whereas decreased significantly for wheat. These findings provide insights on the water demand of the four crops, and thus are meaningful for decision-makers.

Cities at prefectural level(area cities) are not only high-speed economic developing areas, but also the key areas of land supply, reorganization and active transaction. Five variables such as area of land transfer, average land prices, GDP, growth rate of GDP, and fixed assets investment are used to develop a self-organizing feature map(SOFM) artificial neural network model. The results show that 282 area cities in China are divided into the four categories: developed area of high land prices, developed area of low land prices, underdeveloped area of high land prices, underdeveloped areas of low land prices. According to the results, the characteristics of each region are analyzed and the current development situation is discussed. Classification results match the objective reality very well, indicating SOFM-based classification method is an alternative approach in research of socio-economic development.

By using a modified emergy-based ecological footprint model, the authors accounted and analyzed the changing processes of the per capita consumption ecological footprint and local output ecological carrying capacity in Guangdong Province from 1978 to 2006. Growth trend was found in both processes, and the growth rate of the output carrying capacity was higher than consumption ecological footprint. The authors researched the structural characteristics and trends of consumption ecological footprint and local output ecological carrying capacity accordingto the different accounts of biological resources, energy resources, industry, human resource and trade respectively. The growth of the non-biological resources consumption footprint was the most important part of the growth of total consumption, while the growth of product output carrying capacity also benefited from the second major industrial development. Extensive trade and the continuous improvement of human resources since reform and opening up policy were the major reasons for continuous improvement of carrying capacity in Guangdong Province.

Under the support of GIS technique, the authors evaluated the capital stocks of the ecosystem's soil conservation with a case study in Beijing, Tianjin and Hebei Province. On the basis of the data of daily rainfall, NDVI, DEM, LUCC, and soil types of the study area, according to the Revised Universal Soil Loss Equation, the maximum amount of potential soil erosion and its actual quantities were calculated. The deviation between them was the conserved amount, which was converted into capital stocks with the methods of Shadow Price and Market Price. The result show that the value of the soil conservation in the study area was about 176.832 billion RMB yuan, which approximately equaled to the 21.41 percent of the total GDP in 2000, of which the preserved nutrient was about 167.797 billion RMB yuan and the alluvial bedload was 9.035 billion RMB yuan respectively. The contributing percent of individual ecosystems was as follows: the forest occupies 44.34%, cropland 44.91%, grassland and water 10.75%.

Recurrence quantification analysis (RQA) and cross recurrence plot (CRP) are used to analyze the relationship between the normalized difference vegetation index (NDVI) and climatic factors in the Southwest and Southeast Monsoon Regions in Yunnan longitudinal range-gorge region (LRGR). Results show that in the Southwest Monsoon Region the NDVI series has a higher certainty than that in the Southeast Monsoon Region. The coupling relationships of NDVI series in the two regions are between a completely random series and a deterministic series with noise. The two NDVI series have low synchronization in general, especially in the period 1987?1989 and 1991?1993. The NDVI series of the Southwest Monsoon Region lags behind that of the Southeast Monsoon Region. According to the CRP indices, the similarity, persistence time and stability of the coupling relationship between NDVI and precipitation series in the Southwest Monsoon Region are better than that of the Southeast Monsoon Region. Both the NDVI series and temperature series have similar trends. However, the difference of the CRP indices between the two regions is smaller. It is suggested that the interaction pattern of NDVI and temperatures has a smaller difference between the two regions.

Concept model of regional ecological risk is built for the characteristics of ecosystems alongside Qinghai-Xizang highway and railway based on MLP (Multilayer percetron) model. Seven indices such as snow hazard, drought hazard and landslide are selected to evaluate the integrated ecological risk of the ecosystems alongside Qinghai-Xizang highway and railway. Results show that Qaidam mountain desert region has the highest average ecological risk value (4.2585), followed by Guoluonaqu alpine scrub meadow region (2.7640) and Qinghai and Qilian mountain steppe region (2.7335) among these ecosystems alongside Qinghai-Xizang highway and railway through six natural regions. As far as land cover types are concerned, the top three ecological risk values appear in the needle-leaved forest (4.3096), desert ecosystem (4.1174) and land without vegetation (3.6182) respectively, which are higher than those in the other seven land cover types in the study site. Although the risk values are influenced by natural factors and human activities, they are more strongly controlled by natural factors. According to the ecological risk characteristics, these ecosystems alongside Qinghai-Xizang highway and railway are subdivided to four subregions, including Chaidam basin high risk region, Xidatan to Dangxiong moderate risk region, and east Qinghai, Qilian and south Qinghai slight risk regions.