With the rapid development of the social economy and the continuing aggravation of climate change in China, the problem of ecosystem degradation has become increasingly prominent. To ensure the stability and sustainable development of the ecosystem, it is urgent to strengthen the ability of ecological monitoring to recognize the ecological quality and changing trends of ecosystems in China. Carrying out ecological quality monitoring is the basis for studying the changes in ecosystem functions, ecosystem restoration, biodiversity protection and ecological assessment under the influence of climate change and human activities. There is also a major strategic demand for implementing national- scale ecological environment monitoring, ensuring national ecological security and promoting the construction of ecological civilization.

Trans-scale and standardized networked observations are receiving increasingly attention in global change and ecosystem research. The National Ecosystem Observing Network (NEON), the European Integrated Greenhouse Gas Observing System (ICOS) and the Australian Terrestrial Ecosystem Research Network (TERN) are at the interna-tional forefront in the development of global ecosystem observation networks. Constantly enriching and updating the observational technologies, by emphasizing standardized network observation and strengthening the construction of observation systems from station to regional and even national scales, has become the developmental trend of international ecosystem observation and research networks (Novick et al, 2018; Teeri and Raven, 2002). Through the selection of industrial sector stations in China, the Ministry of Science and Technology led the establishment of the National Ecosystem Observation and Research Network in 2006 (Wang et al, 2015). However, there are significant differences in indicator systems, technical means and data specifications among the stations. There is also a lack of effective station-to-region integration of the observation technology. Therefore, it is urgent to develop multi-level and standardized integrated ecosystem monitoring standards and technical specifications to meet the scientific and technological needs of long-term ecological environmental monitoring and dynamic ecological quality assessment at the national level.

Dryland ecosystem is an ecosystem type that supports rare plant communities in an environment of sparse precipitation, strong evaporation and extreme drought. It is widely distributed throughout the whole biosphere and is an important subsystem of the terrestrial ecosystem (China Desert Ecosystem Functions and Research Team, 2017). It is also the regional space associated with national and regional ecological security. After years of development, China's long-term observation of dryland ecosystems has accumulated abundant data on ecological factors, ecological structure, function and biodiversity factors, but has directed only limited efforts toward research on multi-level and standardized comprehensive assessment of the dryland ecosystem. Following a long period of continuous over-exploitation and utilization, the dryland ecosystem has begun to suffer from structural destruction to functional disorder, resulting in regional shortages of water resources, wind erosion and desertification, and loss of biodiversity, all of which pose serious threats to social stability and ecological security in the arid areas of China (China Desert Ecosystem Functions and Research Team, 2017). Therefore, it is urgent and necessary to carry out ecological quality monitoring of dryland ecosystems by comprehensively applying scientific and technical methods to continuously monitor the components, structures and functions of dryland ecosystems at different scales, to obtain multi-level and high-precision information and to evaluate the quality status and changing trends of China’s dryland ecosystems.

In sum, this study aims to: 1) clarify the concept of ecological quality of dryland ecosystems; 2) determine the multiple indicators reflecting the ecological quality and establish the ecological quality indicator system for dryland ecosystems; 3) define the monitoring techniques and methods for each indicator in the indicator system; and 4) establish an assessment model for evaluating ecological quality to comprehensively evaluate the overall ecological quality of dryland ecosystems.

With the rapid development of long-term observational stations and station networks for dryland ecosystems, the number of desert ecological stations in China will reach 47 by 2020, including 6 stations in hyper-arid areas, 9 stations in arid areas, 5 stations in semi-arid areas, 6 stations in sub-humid arid areas, 5 stations in alpine areas on the Qinghai-Tibet Plateau and 16 stations in other special environmental areas (Cui et al, 2017). Relying on the research network of dryland ecosystem positions in China, this research group has carried out the ecological quality monitoring of dryland ecosystems at Kumtag Station located in Dunhuang County in hyper-arid region, Dengkou Station in Bayannaoer in an arid region and Zhenglanqi Station of Duolun County in a semi-arid region.

The dryland ecosystem is a biological community composed of xerophilic and super-xerophilic small trees, shrubs, subshrubs and dwarf subshrubs, as well as the animals and microorganisms that have adapted to it. It is also a dynamic system that allows for material circulation and energy flow together within its habitat. The dryland ecosystem habitat is characterized by sparse precipitation, a dry climate, abundant wind and sand, and sparse vegetation. The dryland ecosystem is the most fragile ecosystem in the land surface process and a representative ecosystem in the arid areas of northwest China, with a unique structure and function. Due to the continuous long-term over-exploitation and utilization of the dryland ecosystem, it has begun to develop structural destruction and functional disorder, resulting in regional shortages of water resources, wind erosion and desertification and the loss of biodiversity, which now pose serious threats to social stability and ecological security in the arid areas of China (China Desert Ecosystem Functions and Research Team, 2017).

Ecological quality refers to the comprehensive characteristics of the elements, structures and functions of the ecosystem within a certain space-time range. It is mainly reflected by the state, production capacity, structure and function stability, resilience to interference and ability of the ecosystem to recover from disruptions, and shows the advantages and disadvantages of the ecosystem's ability to maintain its nature, stability and self-organization. The ecological quality emphasizes the holistic consideration and quantitative scientific assessment of the structure and function of the ecosystem, and directly serves the restoration of the ecosystem, the protection of biodiversity and the establishment of ecological compensation mechanisms.

With the rapid development of various space technologies, positioning observation technologies and computer technologies, ecological quality monitoring is able to adopt scientific, comparable and mature technical methods to monitor ecosystems at different scales and to obtain multi-level and high-precision information. Long-term spatial information of the dryland ecosystem is obtained by using top-down remote sensing technology, geographic information technology and model simulation technology, as well as bottom-up field observation, field investigation and humanistic empirical investigation (Table 1). Based on the ecosystem observation and research network, the data integration analysis is realized. Through multi-source data fusion, scale conversion and satellite-air-ground integration for mutual data verification, the quality status and changes of the ecosystem are assessed, providing important technical support for the protection and restoration of the regional ecosystem. The ecological quality monitoring of the dryland ecosystem includes both remote sensing monitoring and ground monitoring (Liu, 2016).

An assessment of ecosystem quality depends on the representative indicators for ecosystem vigor, organization, and resilience (Costanza, 1992). However, as each indicator can be quantified using different units, the values of each indicator need to be normalized. In order to avoid magnifying the indicators when they have different units, it is necessary to neutralize the order of magnitude:

${{M}_{i}}=({{m}_{i}}-{{m}_{\min }})/({{m}_{\max }}-{{m}_{\min }})$ (1)

Where m_i is the value of the indicator datasets; and m_min and m_max refer to the minimum value and maximum value of the indicator datasets, respectively.

Both model indicators affect the assessed quality of the ecosystem and they describe different aspects of the ecosystem structure and features. The weights of these different aspects were set according to the Delphi method (Dalkey and Helmer, 1963) and the analytic hierarchy process (AHP) (Saaty, 1977). Firstly, the ecosystem quality indicators were decomposed into different hierarchy levels. Secondly, the correlations among the indicators were weighted based on the opinions of experts. Lastly, the judgment matrices were formulated and the importance of the sub-level indicators for the total indicator and the weights of the indicators were calculated (French et al., 1991). The formula used to calculate the total ecosystem quality is as follows:

$T{{Q}_{i}}=\sum\limits_{j=1}^{n}{{{\beta }_{ij}}{{M}_{ij}}}$ (2)

where TQ_i is the ecosystem quality score of the i^th assessment unit; β_ij is the weight of the j^th indicator of the i^th unit; and M_ij is the normalized value of the j^th indicator of the i^th unit. The ecosystem quality grades were classified by the TQ of both assessment units.

In this study, ecosystem quality is modeled at two scales: the regional scale and the station-scale when ecosystem quality was evaluated. The indicators in the model would correspond to the specified scales.

This study presents the scientific conceptual framework for monitoring and evaluating the ecological quality of the dryland ecosystem, which includes the concept of ecological quality, an indicator system, monitoring technologies and assessment models. The indicator system consists of 3 indices at the regional scale (land use cover change, net primary production and soil organic carbon) and 5 indices at the station scale (vegetation coverage, species number, plant biomass, net primary productivity of plants, soil carbon content). These indices can be monitored directly by satellite UAV, ground sensor networks and ground measurements at multiple sites. The comprehensive quality of the dryland ecosystem is evaluated by an assessment model that integrates the composite indicators and their weights. Integrated monitoring and assessment planning for ecological quality of dryland could become the tool for dryland ecosystem management to deliver multiple environmental and socio- economic objectives, including the pursuit of the Sustainable Development Goals in arid and semiarid area.

We especially thank professor Wang Shaoqiang and Lu Qi for their support and valuable contribution and discussions. We also gratefully thank professor Feng Yiming for the help on Fig. 1.