National Flash Flood Disasters Investigation and Evaluation project is the largest non-engineering projects in water conservancy industry since the establishment of new China in 1949. It is also the largest scale of general census on disasters background in flood management and mitigation fields. The whole project lasted for 4 years, covering 30 provinces, 305 cities and 2138 counties, with a total land area of 7.55 million km2 and a population of nearly 900 million. Through general census, on-site investigation, field measurement, hydrological analysis and calculation, and comprehensive evaluation methods, the spatial distribution, human settlement, underground situations, impacts, social and economic situations, hazard zoning, warning indicators and historical situations of flash flood disasters were collected. The storm flood characters in mountainous areas were also analyzed. The flood control ability of selected villages were assessed and the critical rainfall index of these villages were obtained. The hazard zones were finally identified, all of which provided a strong information support for flash flood early-warning and forecast and residential safety transfer. We systematically introduced the key focuses on the investigation and evaluation project of national flash flood disasters, made a general review on the collection of data and information, summarized thousands of investigation results and elements during this huge project. We also discussed the spatial pattern of these elements. Based on these survey data, the characteristics of flash flood disaster prevention areas, the human settlement features and spatial pattern of storm flood were further analyzed. Finally, flash flood prevention areas, population distribution, flash flood warning ability and historical flash flood disaster events were discussed. It was found that the national flash flood prevention areas, human settlement, historical flash flood events and warning ability appeared to be spatially consistent. They were mainly distributed along the transitional zone of Qinghai-Tibet plateau and Sichuan basin, the borders of Sichuan and Yunnan provinces, the Loess plateau zones, the Eastern coastal areas and the North China areas. Meanwhile, future application and analysis on diversified utilization of investigation and evaluation results of national flash flood disasters were proposed, providing a solid data foundation for flash flood monitoring and warning system, disaster management and mitigation researches, a better platform of technological promotions, in both flood management departments and other relevant fields.
Mountain flood is difficult to predict and distributes differently. It is one of the major natural disasters in China. Analysis on the distribution characteristics of mountain flood is helpful to disaster prevention. This paper takes historical mountain flood disaster points as based-data and analyzes the relation between disaster points and altitude, rainfall (H6-50), altitude standard deviation in different geomorphologic zones. Results shows that among the 6 geomorphologic zones, the historical mountain flood disasters mainly distribute in zone II and zone V, nearly 60% of the whole disasters. Also, density of mountain flood disaster reaches its high point with rainfall (H6-50) ranging from 240 to 280 mm, altitude standard smaller than 30 m. With the increase of rainfall (H6-50), the mountain flood density tends to increase first and then decline, taking 280 mm as a turning point. What’s more, the mountain flood density reaches its climax with altitude standard smaller than 30 m. In zone I, mountain flood disaster density reaches its high point with rainfall (H6-50) ranging from 240 to 300 mm, altitude from 60 to 120 m, altitude standard smaller than 30 m. For zone II, mountain flood disaster density reaches its high point when rainfall (H6-50) ranging from 150 to 270 mm, altitude from 10 to 50 m, altitude standard smaller than 30 m. Mountain flood disaster in zone V mainly distribute in the area where rainfall (H6-50) higher than 120 mm, altitude under 600 m, altitude standard smaller than 50 m. It is clear that taking the geomorphologic zoning as an analysis unit is much better than administrative boundary.
China is a geological disaster-prone country. Frequent geological disasters cause a lot of casualties, and lead to serious economic losses and the damage of ecological environment. Therefore, it is of great scientific significance and practical value to acquire the occurrence characteristics and the loss status of geological disasters. Based on the "National Geological Disaster Bulletin" issued by Chinese Institute of Geo-environmental Monitoring, we analyze spatial-temporal changes and influencing factors of geological disasters in China from 2005 to 2016 in five aspects: the change of occurring types of geological disaster, the spatial distribution of geological disaster, the cause of geological disaster, the loss caused by geologic disaster and the avoidance situation of geological disasters. The results show that: (1) Landslide and collapse are the main types of geologic disaster, which account for 70% and 10% of total geologic disasters, respectively. The total number of disasters is obviously reduced, especially the collapse is obviously reduced; meanwhile, the number of the landslide disasters has some reduction. (2) For the spatial distribution, geologic disasters mainly distributed in Hunan province and Sichuan province and so on, and they are also the provinces with the highest economic losses. Sichuan is also the province with the largest number of deaths and missing. Serious geological disasters mainly occur in Sichuan province, then Hunan province, Guizhou province and Anhui province. (3) For the cause of disaster situation, the geological disasters caused by natural factors decreased from 96.6% to 92.0%. Compared to natural causes, the proportion of geological disasters caused by human is increased, which is about 0.5% annually through regression equation analysis. With successful transformation of Chinese economy and the method improvement of disaster prediction and prevention, geologic disaster caused by human factors may gradually be stable and few slowly. (4) For the event of disaster losses, direct economic losses have reduced from 4.09 billion Yuan to 3.17 billion Yuan, and the number of dead and missing people has also significantly reduced, about 75 persons per year. Geologic disasters of super-huge types only account for 0.5% of total number, which cause 25.7% of total number of injury and dead people and 47.7% of total direct economic loss. (5) In terms of the avoidance of geological disasters, the percentage of the avoidance number of geological disasters to the total number of geological disasters increased from 2.8% in 2005 to 7.0% in 2016, and the percentage of avoidance economic losses to direct economic losses was increased from 9.3% in 2005 to 22.4% in 2016. The percentages of the avoidance number of geological disasters and avoidance economic losses are about 0.7% and 1.5% annually perceptively using regression equation. Through the analysis of the long term dynamic monitoring results of Chinese geological disasters, it is found that Chinese disaster prevention and mitigation has made remarkable progress.
The economic losses caused by torrential disaster in China are increasing. Housing damage plays an important role in economic losses and casualties. Based on the data provided by the survey and evaluation of torrent disaster, we analyzed the temporal and spatial distribution features of house exposure, constructed the damage risk assessment model of torrent disaster, analyzed the causes of housing damage risk and explored the reasons. The results showed that: (1) The houses in China rural area are mainly one-layer building, and distribution of one-layer dominated county is roughly the same with the boundary of the third terrain ladder of China; brick-concrete structure is the main housing structure in mountainous rural areas, followed with brick-wood structure, steel-concrete structure and other structures are the least. (2) Overall, the housing vulnerability is high in the northwest and low in the southeast, while the housing damage risk is high in the east and low in the west. The areas with extremely high damage risk are concentrated in the Liaodong Peninsula, Shandong Peninsula, Hainan Island and the Southeastern coastal areas and Yanshan-Taihang Mountains. The areas with high damage risk are banded or agglomerated, which are mainly distributed in the Taihang Mountains and middle-lower reaches of Yangtze River. (3) The high housing damage risk in Shandong, Shanxi and Hebei province are more relied on housing structure types. The houses with high damage risk which are mainly due to torrent strength are located within three stripes: Yanshan-Taihang Mountains Belt, Zhejiang-Fujian Coastal Hilly Belt, Guangdong-Guangxi Coastal Belt.
The region of loess plateau is one of the ecologically fragile areas in China. It is one of the slippery strata of which landslides often developed. The development of the loess landslides is the result of the combined effect of many factors such as the disaster environment, the disaster factor and the disaster bearing body. The selection of topographic factors which are important environmental elements of the disasters is the basis of the study of loess landslide risk. We take Gangu county as the study area, which is one of the frequent occurrence of loess landslide hazards in Gansu province. According to the basic conditions of landform, suitability analysis of topographic factors in the study area was carried out with models of sensitivity index (SI), certainty factor (CF) and correlation coefficient (CC). The following conclusions were obtained: Based on the models of SI, CF and CC, the topographic factors suitable for the loess landslide hazard research of this area are: slope, slope of slope (SS), slope shape and surface roughness (SR). The selection of disaster-inducing factors with the methods of SI and CF is merely based on the analysis of the relationship between the single factor and the landslides. We ignore the correlation between the topographic factors. The experimental results indicated that the area with poor stability in the study area has a good correspondence with the number of landslide hazard distributions. After deeply analyzing the relationship between the landslides and the grade range of the topographic factors, it is found that the grade range of the topographic factors has an important influence on the landslide risk assessment of the study area. This is one of the main reasons of the differences in some areas. Based on field investigation, the river-cutting depth, drainage density and human engineering activities are also important topographic factors and have key control effect on the landslides of this study area.
In this paper, we use the information volume model to study the hazard degree of mountain torrent disaster in hilly area of Hunan province. The greatest significance of the information volume model is that it can find the best combination of factors from various factors that affect the occurrence of mountain torrents. Hunan Province is one of the most serious mountain torrent disaster provinces in China. In order to study its hazard degree of the mountain torrent disaster, we divide Hunan Province into several small watersheds to evaluate the hazard of mountain torrent disaster using the data of historical flash flood disaster points spanning years from 1955 to 2015. Information volume model was established to calculate the information of the six factors: elevation, slope, relief, land cover, soil types and rainfall, respectively from the consideration of three aspects including the terrain, the underlying surface and the rainfall. The information volume of six factors were calculated, respectively. Based on the value of information volume of these factors, we obtained the combination of factors with the biggest influence of flash flood disasters. Through calculating the total value of information volume for all small watersheds in hilly areas of Hunan Province, we classified the information volume into five types associated with different dangerous levels. The results show that: the most significant contribution to the flash flood is the artificial surface of land cover type, with a information volume of 1.771, followed by the types with relief degree less than 30 m as well as the clay soil type (both at a value of 1.331). The mountainous torrent disaster for Hunan hilly area are likely occurred in areas with slope lower than 10°, elevation lower than 100 m, relief lower than 30 m, of which the land cover is artificial surface, the soil type is clay and annual mean rainfall is between 1584.3~1662.0 mm. Statistics of each level of dangerous areas show that the second-high and the third-high hazard types have the largest area in the mountain areas of Hunan, accounting for 26.59% and 26.63%of the total mountain areas, respectively. Area percentage of the fourth-high risk type is 20.89%, and that of the first and fifth-high hazard types is 13.89% and 11.99% respectively. In the hilly areas of Hunan province, cities with higher hazard levels are Yongzhou City, Chenzhou City, Zhuzhou City, Yueyang City, Loudi City and the eastern part of Changsha. In this study, 90% of the mountain torrents (1243 mountain torrents) were selected randomly from 1381 mountain torrents disaster spots in Hunan Province, and 10% of them were used to verify the hazard assessment results. The verification of confusion matrix demonstrated that the accuracy rate of this model is 75.36%, indicating a basically credible results of the spatial distribution of hazard degree estimated in this study. The mountain torrent disaster system established in this study still needs to be improved. The selection of factors and models, as well as the quantification of human activity factors should be considered in the further study.
Hazard assessment is important for the prevention and mitigation of debris flow disaster. This study takes Sichuan Province as the research area. Based on the DEM data, we realize the demarcation of the small watershed in Sichuan Province by extracting the direction of water flow and calculating the accumulation of the flow confluence. Based on the collected information of debris flow watershed, we selected the watershed elevation difference and watershed area as the indicators. The identification model of potential debris flow watershed based on the energy condition was built by analyzing the hazard-formative environment and characteristics of debris flow hazards. A total of 7798 small watersheds with the required energy conditions for debris flow occurring were identified by utilizing the established model among all the demarcated small watersheds, which is 31.1×104 km2 accounting for 64.18% of the total area of Sichuan Province. The indicator system of debris flow hazard risk assessment and the extension matter-element model were established from the energy condition of the debris flow occurring, the condition of the loose solid materials, the precipitation condition and the condition of human activity. These determine the weights of the assessment factors, of dividing the grade of the hazard risk, by which it classifies the hazard risk degree of small watershed debris flow. The number of the moderate, high and very high hazard degree is 1946, 1725 and 1002, with an area of 9.1×104 km2, 7.7×104 km2 and 3.4×104 km2, respectively. The total area of moderate hazard areas is 20.2×104 km2, accounting for 41.67% of the total area of Sichuan Province. Finally, the analyses were made for the reliability of assessment results and the distribution of the different hazard degree of debris flow areas in different municipal administrative districts and the major river valleys. All the known small watershed of very high hazard degree are identified as debris flow watersheds. The 896 watersheds of moderate hazard degree do not belong to the identified debris flow and 1233 watersheds of high hazard degree do not belong to the identified debris flow, either. They are the key area for disaster prevention and reduction in Sichuan province in the next few years. The results of the analyses have the great theoretical and practical significance for enhancing the debris flow identification, the prevention and mitigation of regional debris flow disaster. The sustainable development of mountainous areas and also the theory of the risk assessment of debris flow hazard.
Taking Huoxi Coal Mine Area in Shanxi Province as the research area, we conducted numerical modeling and quantitative evaluation of landslide susceptibility using remote sensing and GIS technology. Based on the DEM with spatial resolution of 30 m × 30 m, five topographical parameters were derived: elevation, slope angle, slope aspect, plan curvature and profile curvature. Stratigraphic lithology was digitized based on the geological maps from Department of Geological Survey in 1:50 000 scale. Fault network, drainage network and road were digitized based on the geological maps and other thematic maps from Department of Land Resource in 1:50 000 scale. Then, buffer for faults, drainage, and road were done. Mining disturbance were digitized based on the planning maps of coal resources. If the point falls in the mine area, it is proved to be disturbed by the mining disturbance, otherwise is not affected. NDVI and land-use types interpreted and computed the Landsat TM images. Landslide data was collected by Bureau of Land and Resources and it is represented by the X, Y coordinates of its central point. Then, the correlation characteristics among evaluation factors and the spatial distribution of landslides were acquired by using remote sensing technology and GIS spatial analysis method. Repeated 5-fold cross validation method was adopted in this research and the landslide/non-landslide datasets were randomly split into a ratio of 80:20 for training and validating models. Based on the methods of the 5-fold cross-validation and the fitting accuracy to the constructed the landslide susceptibility assessment model-Radial Basis Function - Support Vector Machine (RBF-SVM), the precision of the models was quantitatively assessed. We calculated the importance of each evaluation factor in the RBF-SVM model. Meanwhile, we obtained landslide susceptibility map of Huoxi Coal Mine Area based on the RBF-SVM model. The landslide susceptibility of Huoxi Coal Mine Area was divided into four scales referencing the quantile law: low (0-0.02), medium (0.02-0.1), high (0.1-0.85) and very high (0.85-1) probability of landslide. The results show that: (1) the fitting accuracy was 87.22% in the modeling phase and 70.12% in the validation phase, respectively, for the RBF-SVM model; (2) it indicated that lithology, distance from road, slope aspect, elevation and land-use types have contribution to each model. Therefore, these five factors are most suitable conditioning factors for landslide susceptibility mapping in this area. Mining disturbance factors have little contribution to the model. The mining method in this area is underground mining and the mining depth is very deep affecting the stability of the slopes. (3) The number of landslides points in the very high region was 316, which account for 93.49% of the total number of landslides points and 50.99% of the total area. This study obtained the spatial distribution characteristics of the Huoxi Coalfield geological disasters and the quantitative evaluation of landslide susceptibility. It provides reference for the investigation about artificial slope in the research area monitoring the rational mining coal resources. It will also provide the reference for the related research in other similar coal region and management work.
The Newmark displacement model is a common physically based model for earthquake induced landslide susceptibility mapping. The machine learning model is one of the statistical methods and is increasingly applied to landslide susceptibility mapping in recent years. The main purpose of this paper is to combine the Newmark displacement method with machine learning model for developing a mechanism-based earthquake-induced landslide susceptibility model and improving the predictive accuracy. The support vector machine (SVM) method was selected and eight thematic data layers, including landslide inventory, topographic relief, slope, curvature, Newmark displacement value, distance from faults, distance from drainages and distance from roads, were prepared in GIS. A total of 1,900 landslides were subsequently randomly divided into two subsets: a training subset comprising 70% of the landslides and a validation subset containing the remaining 30%. The model is then applied to Wenchuan County, which was one of the most severely affected areas during the May 12, 2008 (Mw 7.9) Wenchuan earthquake in China. The model performance was evaluated using the receiver operation characteristic (ROC) curve and the model estimation uncertainty in comparison with the other two statistical methods: frequency ratio (FR) bivariate statistical model and the logistic regression (LR) multivariate statistical model. The results show that five variables including topographic relief, Newmark displacement value, distance from faults, distance from drainages and distance from roads are finally selected based on a significance level of 0.05 and the multi-collinearity detection. The values of the area under the ROC curve (AUC) demonstrate that the SVM model exhibited the highest accuracy for the training and validation data sets with AUC values of 0.876 and 0.851, respectively, followed by the LR model (AUC values of 0.836 and 0.842 for training and validation, respectively) and FR model (AUC values of 0.844 and 0.808 for training and validation, respectively). For the evaluation of the model prediction uncertainties, the pixels classified as high susceptibility (probability ≥ 0.75) and low susceptibility (probability < 0.25) are more valuable and practical, both of which have high reliability to determine whether the location is stable. The prediction variation is low for pixels classified as high susceptibility and low susceptibility (with an average of the standard deviation less than 0.05), indicating that the SVM and LR models consistently identified these pixels as stable or unstable. Furthermore, these high and low susceptibility pixels account for about 70% of all pixels for the SVM model, which is about 25% higher than that of LR model. It means that the SVM model performs better than the LR model in these high reliability pixels. For the pixels classified as intermediate susceptibility (probability 0.25 ~ 0.75), the standard deviation of the predictive probability of the SVM model is about 0.09, which is larger than that of the LR model. It indicates that the SVM model exhibited larger uncertainty than the LR model in these intermediate susceptibility pixels. However, it is difficult to determine whether these pixels are stable or not. Also, the pixels with intermediate susceptibility only account for about 30% of the total samples for the SVM model. In general, the SVM model combined with the Newmark displacement method outperform the LR model in accuracy and uncertainty evaluation. The proposed method was applied to produce the earthquake-triggered landslide susceptibility map of Wenchuan County. The comparison of landslide susceptibility map and actual landslide distribution showed that the high susceptibility areas can account for about 80.4% of the actual landslides. This indicates that the combination of support vector machine and the Newmark displacement method has a higher predictive value. The proposed method can potentially help risk assessment and effective management of landslides caused by earthquakes.
Flash flood disaster ranks top among the natural hazards in China due to its frequent occurrence and high mortality. It has posed a severe threat to national public safety and water security. The optimal design of hydrometric network helps to capture the spatial and temporal variability of regional storm events and flash flood regimes, which can significantly improve the accuracy of flash flood warning and strengthen the flood defense capacity in small and medium catchments. However, the quantitative research on the optimal design of hydrometric network is relatively rare with consideration of flash flood disaster prevention and control. In this study, a flash flood warning-oriented method for the optimal design of hydrometric network is proposed, and Shunchang County in Fujian Province which is frequently subject to flash flood disasters is selected as a case study. Specifically, the optimal design method for rainfall station includes the Cone method and the correlation analysis method, and should further consider the characteristics of historical storm events and flash flood disasters, and local social and economic situations. The optimal design of water stage station should comprehensively consider the occurrence frequency and impact extent of historical flash flood disasters, flash flood warning requirements, and local social and economic situations. The current densities are 37 km2 per station and 76 km2 per station for rainfall stations and water stage stations, respectively, while most of the stations are distributed along the main streams in the plain, and the monitoring stations are insufficient especially in the key prevention districts for flash flood disasters. Thus, the current monitoring capacity and early warning capacity are rather weak in Shunchang County, which significantly weakens the warning and forecasting capacities of local flash flood disasters. For addressing those above-mentioned problems, the current designs of rainfall station and water stage station are quantitatively analyzed and optimized according to the spatial and temporal characteristics of storm events, the historical flash flood disasters and the forecasting requirements of flash floods. The results show that three new rainfall stations and three new water stage stations are suggested to be established in Shunchang County, in which one water stage station monitors rainfall process meanwhile. The three new rainfall stations are evenly distributed along the densely populated mountainous tributaries with rare rainfall monitoring stations and relatively weak flood defense capacity. Their establishments can improve the forecasting accuracy of local flash floods. The three new water stage stations are evenly distributed along the densely populated mountainous tributaries with relatively weak flood defense capacity, and their establishments can guarantee the forecasting requirements of flash flood disasters, and the safety of downstream villages and water infrastructures. After the optimal design of the hydrometric network in Shunchang County, the density of rainfall station and water stage station arrive at 34 km2 per station (31 rainfall stations) and 68 km2 per station (29 water stage stations), respectively. This study is expected to provide scientific references and technical guidance for the robust and quantitative design of hydrometric network in regions frequently hit by flash floods.
Flash flood early warning is an important non-structural measure for flash flood prevention in China. Also, rainfall threshold is the key for flash flood early warning. At present, the method of calculating rainfall thresholds need a large amount of meteorological and hydrological data. Meanwhile, building the hydrological model and calibrating parameters are difficult, which are not suitable for the flood control personnel. A simple and easy method, using flood peak modulus to calculate rainfall threshold, was proposed in this study based on national flash flood investigation and evaluation results. Rational equation is the basis of the calculation method. The flood peak modulus in rational equation is expressed as a ratio of the flow to watershed area. Then, the critical rainfall formula is obtained. Using the Manning formula and national investigation and evaluation results, the flood peak modulus on the condition of critical flow is obtained, and the net rainfall is calculated. Based on the research results of the scholars, three aspects about the rainfall loss calculation are considered. They are depression storage, vegetation interception and soil infiltration. The sum of the net rainfall and the rainfall loss is the critical rainfall. Considering the factors such as soil water content of watershed, the rainfall threshold was finally obtained. In order to demonstrate this method, Shuanghe catchment in Suijiang County of Yunnan Province was chosen as the study area, of which the area is 89.12 km2. The calculated concentration time was 5.2 h. Thus, the duration was estimated to be 1 h, 3 h, and 6 h. The results indicated there was a linear correlation between the net rain amount and different rainfall durations. Depression storage and canopy interception was invariant during different rainfall durations, but the infiltration was variable. For 1 h duration, initial infiltration was the main factor to consider; for 3 h, both initial infiltration and mid-term infiltration were considered; and for 6 h, steady infiltration also needed consideration besides the initial and medium-term infiltration. The calculated critical rainfall for 1 h, 3 h, and 6 h were 38.6 mm, 64.8 mm, and 96.9 mm, respectively. Rainfall thresholds of different durations for flash flood early warning were estimated on basis of critical rainfall considering three different soil moisture conditions. Under dry soil moisture condition, large rainfall loss leads to a large rainfall threshold; under wet soil moisture condition, contrary to the dry condition, small rainfall threshold is caused by small rainfall loss; and under medium soil moisture condition, the rainfall threshold was in medium. The rationality analysis on critical runoff, rainfall losses, and rainfall threshold was carried out in the end of this study. The results showed the calculated rainfall threshold by flood peak modulus method was approximated to the rainfall threshold obtained from national flash flood investigation and evaluation project. Also, they are in accord with the observed rainfall during flash flood events. Thus, the calculated results are reasonable in this study. This study provided a quick and convenient way of calculating rainfall threshold of flash flood warning for the grass root staffs and offered technical support for estimating rainfall threshold correctly.
"Natural disaster" is the phenomenon of the losses of life and property, which is caused by the interaction of human society and natural environment. It’s also the product of the disaster environment, disaster-causing factors and disaster-bearing body. In order to study the processes, mechanisms and impacts of natural disasters as well as the reduction of the losses caused by natural disasters, it is necessary to conduct surveys of basic data and natural disaster events on a large scale of which the authenticity and consistency are much significant for ensuring the reliability and validity of the research results. However, the large number of organizations and investigators participating in the survey and evaluation process, large regional differences and large spatial scale create challenges in data quality control and validating the consistency of data from various survey units. To ensure the correctness and consistency of the data, it is necessary to carry out manual inspection. However, for the massive survey data, it is unrealistic to totally rely on manual work to effectively identify the abnormities. As a result, we design a multi-scale anomaly detection method for natural disaster survey data by using the single-element detection method of outliers based on normal distribution and spatial clustering method of Anselin Local Moran's I to detect the abnormal values and abnormal spatial distribution patterns of the massive survey data. It can effectively extracts the abnormalities and abnormal investigation units at all levels of scale and gains the reasons for abnormal data. It provides the support for the manual checking of survey data. In this paper, taking the project of flash flood disaster investigation and evaluation in mainland of China as an example, this method is used to audit the events of historical flash flood disaster and the areas of the towns which are in the prevention zones. Also, it quickly extract the anomaly units of flash flood disaster point density and township units with exceptional area values. Further analysis found that the reasons for these abnormalities were due to the inconsistency of filling methods, unit errors, and repetition of records and so on. The method resolved the inconsistency in massive amounts of flash flood survey data. This method is an effective approach of checking the quality of various other large-scale disaster datasets. Although the data validation approach used in this study is very effective, there are still some problems, i.e. the outlier checking only considers the outliers between survey units based on the administrative divisions. Regions are not divided according to their economic development and natural conditions. Finally, we analyze the applicable conditions of this method in the large-scale natural disaster investigations.
Under the context of global climate change, the heavy flood caused by the snow melting (glacier melting) as well as heavy rainfall in the high altitude mountainous areas in Xinjiang Uygur Autonomous Region was becoming more unpredictable. Therefore, clarifying the relationship between the temperature and the rainfall types is the prerequisite step to predict the flood effectively in these mountainous regions. Fortunately, the approach of rainfall and snowfall separation in mountainous regions is capable of determining the temperature conditions which may cause the heavy flood. It is also able to provide important and scientific references to the accurately prediction for the heavy flood in the mountainous regions. In this study, temperature and precipitation data were collected from ground-based meteorological stations located in different altitude in a case study area: the Tizinafu River Basin in Kunlun Mountains. This study was conducted on a daily basis during 2012 to 2016. The MODIS10A2 snow cover data with 8-day temporal resolution were also applied as the valid reference data. For the purpose of rainfall and snow separation, we adopted the temperature integral and probability statistics methods to analyze the temperature conditions for different rainfall types in the research region. The remote sensing snow cover data combined with the average temperature over the latest past few years are used to determine the different temperature conditions with different precipitation patterns. The results were summarized as follows. If the maximum temperature and accumulated temperature reaches 20.91 ?C and 51.82 ?C, respectively, the precipitation can be predicted as rainfall in the Momuke station. In contrast, if maximum and accumulated temperature are below 18.13 ?C and 43.69 ?C, respectively, the precipitation can be predicted as snowfall. Similarly, for Kudi station, if the maximum and accumulated temperature reaches 14.51 ?C and 33.17 ?C, respectively, the precipitation can be judged as rainfall. While the precipitation will be recognized as snowfall when the maximum and accumulated temperature are below 13.57 ?C and 31.68 ?C, respectively. In the same way, when the maximum temperature and accumulated temperature in the Xihexiu meteorological station are above 9.43 ?C and 19.53 ?C, respectively, the precipitation will be recognized as rainfall and the precipitation will be recognized as snowfall once maximum temperature and accumulated temperature are below 8.22 ?C and 19.4 ?C, respectively. For validating and evaluating the credibility of this rainfall and snowfall separation method as well as the reasonability of the reference temperature conditions, the meteorological data from the nearby villages of the study catchment were used to assess rainfall and snow separation results. From the results, we can conclude that in different elevation bands, the rainfall and snow separation results are always acceptable with different levels. The precisions are 92.86%, 79.49% and 88.3% in the elevation bands below 2000 m, 2000-3000 m, and 3000 m above sea level, respectively. The results is capable of providing a scientific evidence for monitoring flood types and flood forecasting, which is of great significance and is related to create new water resource management guidelines and planning schemes for local people and decision makers.
