With the environmental deterioration caused by the advance of climate change, soil salinization is a serious and growing global problem. Currently about 7% of the world's land surface is threatened by salinization. China is a country whose soils are severely affected by this problem, which, due to its extensive area, and wide distribution poses a serious threat to regional agricultural development. In this review, we summarize the framework for soil salinization research in China over the past 70 years, assess the weaknesses of existing research in both a domestic and international context, highlight the trends and key findings of global research about saline soils over the past 30 years, and propose six major fields and directions for future research on saline soil.

Based on the palaeoclimatic and palaeoenvironmental evidences of geological history and human history periods, this paper reviews the researches and progresses on the development of the sandy deserts in Xinjiang. It pointed out that the features of tectonic structure in Xinjiang had made both the Tarim Basin and the Junggar Basin being influenced greatly by the foehn effects originated from the planetary wind system of westerly, the East Asian ocean-continental monsoon and the topographical mountain-valley winds. The regional patterns of climate and environment since the Quaternary were characterized by the overall persistent drought accompanied by fluctuations in the secondary scale. Formations of aeolian sediments in the basins and at the margins are a potential response to global climate change, particularly the aridification of the Asian hinterland deduced by the uprising of the Tibetan Plateau and its surrounding highlands. For the question about the formation time of the Taklamakan Desert, because the research methods, objects and information carriers used in previous studies are different, there are many disputes in the academic circles at present. Evidences from aeolian deposits/rocks at the edge and in the hinterland of these sandy deserts and their chronological data indicate that an arid climate and land surface aeolian processes have occurred at the edge of the Tarim Basin and its hinterland areas since the Tertiary period. However, the duration time of these processes at mass scale should have begun after the middle Pleistocene and lasted to the Holocene. Occurrence of dune fields in recent 2000 years in the oasis areas should be greatly influenced by human factors.

Language plays a vital role in the communication, sharing and transmission of information among human beings. Geographical languages are essential for understanding, investigating, representing and propagating geo-spatial information. Geographical languages have developed and evolved gradually with improvements in science, technology and cognitive levels. Concerning the theoretical progress from geographical information ontology, epistemology and linguistic theory, this paper firstly puts forward the concept of a GIS language and discusses its basic characteristics according to changes in the structures, functions and characteristics of geographical languages. This GIS language can be regarded as a system of synthetic digital symbols. It is a comprehensive representation of geographical objects, phenomena and their spatial distributions and dynamic processes. This representation helps us generate a universal perception of geographical space using geographical scenarios or symbols with geometry, statuses, processes, spatio-temporal relationships, semantics and attributes. Furthermore, this paper states that the GIS language represents a new generation of geographical language due to its intrinsic characteristics, structures, functions and systematic content. Based on the aforementioned theoretical foundation, this paper illustrates the pivotal status and contributions of the GIS language from the perspective of geographical researchers. The language of GIS is a new geographical language designed for the current era, with features including spatio-temporal multi-dimension representation, interactive visualization, virtual geographical scenarios, multi-sensor perception and expedient broadcasting via the web. The GIS language is the highest-level geographical language developed to date, integrating semantic definitions, feature extraction, geographical dynamic representation and spatio-temporal factors and unifying the computation of geographical phenomena and objects. The GIS language possesses five important characteristics: abstraction, systematicness, strictness, precision and hierarchy. In summary, the GIS language provides a new means for people to recognize, understand and simulate entire geo-environments. Therefore, exploration of the GIS language’s functions in contemporary geographical developments is becoming increasingly important. Similarly, construction of the conceptual model and scientific systems of the GIS language will promote the development of the disciplines of geography and geographical information sciences. Therefore, this paper investigates the prospects of the GIS language from the perspectives of digital technology, geographical norms, geographical modeling and the disciplinary development of geography.

Sandy desert landscape is a geological product of arid climate and abundant sand materials supply in natural conditions, therefore the formation of sandy desert is an ideal studying object for understanding the interaction between various stratigraphic/epigenetic spheres of the earth system. However, until now, the knowledge about the provenance and formation of these deserts in Xinjiang, Central Asia is diverse and inconsistent, a systematic review is necessary. Ancient aeolian sand sediments in the internal areas of these deserts and their twin dust deposits in the Cenozoic strata surrounding the Xinjiang deserts are direct and indirect clues reflecting the provenance of sand materials and the formation and evolution of deserts. Based on the geochemical, mineralogical and isotopic evidences of desert sands and relevant deposits in the peripheral regions, this paper reviews the research progress on the development of the sandy deserts in Xinjiang, Northwest China. Many proofs proposed that desert sands in Xinjiang were mainly sourced from the ancient pluvial, alluvial and fluvial sediments and were eventually achieved because of the local sand supply. It pointed out that the settings of tectonic structure in Xinjiang had made sediments both in the Tarim Basin and the Junggar Basin being influenced greatly by regional hydrological system and aeolian processes originated from the planetary wind system of westerly, the East Asian winter monsoon and the topographical mountain-valley winds. However, the directions of transportation paths are different between the hydrological dynamics and the aeolian dynamics, which are decided by the slantwise structures of the basins tectonics and the regional atmospheric circulation routes, respectively, and have determined anisotropies of the transport pathway of these sediments delivered by different agents. This results in a geological cycle of detrital sediment transportation in the two large basins and thus largely increases the degree of sediment mixing, alternation and recycle between younger and older sediment sources, as well as the complexity and diversity of sand provenances.

Statistical models using historical data on crop yields and weather to calibrate relatively simple regression equations have been widely and extensively applied in previous studies, and have provided a common alternative to process-based models, which require extensive input data on cultivar, management, and soil conditions. However, very few studies had been conducted to review systematically the previous statistical models for indentifying climate contributions to crop yields. This paper introduces three main statistical methods, i.e., time-series model, cross-section model and panel model, which have been used to identify such issues in the field of agrometeorology. Generally, research spatial scale could be categorized into two types using statistical models, including site scale and regional scale (e.g. global scale, national scale, provincial scale and county scale). Four issues exist in identifying response sensitivity of crop yields to climate change by statistical models. The issues include the extent of spatial and temporal scale, non-climatic trend removal, colinearity existing in climate variables and non-consideration of adaptations. Respective resolutions for the above four issues have been put forward in the section of perspective on the future of statistical models finally.