Non-crossing Quantile Regression Neural Network as a Calibration Tool for Ensemble Weather Forecasts

Despite the maturity of ensemble numerical weather prediction(NWP),the resulting forecasts are still,more often than not,under-dispersed.As such,forecast calibration tools have become popular.Among those tools,quantile regression(QR)is highly competitive in terms of both flexibility and predictive performance.Nevertheless,a long-standing problem of QR is quantile crossing,which greatly limits the interpretability of QR-calibrated forecasts.On this point,this study proposes a non-crossing quantile regression neural network(NCQRNN),for calibrating ensemble NWP forecasts into a set of reliable quantile forecasts without crossing.The overarching design principle of NCQRNN is to add on top of the conventional QRNN structure another hidden layer,which imposes a non-decreasing mapping between the combined output from nodes of the last hidden layer to the nodes of the output layer,through a triangular weight matrix with positive entries.The empirical part of the work considers a solar irradiance case study,in which four years of ensemble irradiance forecasts at seven locations,issued by the European Centre for Medium-Range Weather Forecasts,are calibrated via NCQRNN,as well as via an eclectic mix of benchmarking models,ranging from the naïve climatology to the state-of-the-art deep-learning and other non-crossing models.Formal and stringent forecast verification suggests that the forecasts post-processed via NCQRNN attain the maximum sharpness subject to calibration,amongst all competitors.Furthermore,the proposed conception to resolve quantile crossing is remarkably simple yet general,and thus has broad applicability as it can be integrated with many shallow-and deep-learning-based neural networks.

Spatial and Temporal Distribution Characteristics of Solar Energy Resources in Tibet

The Tibet Plateau is one of the regions with the richest solar energy resources in the world.In the process of achieving carbon neutrality in China,the development and utilization of solar energy resources in the region will play an important role.In this study,the gridded solar resource data with 1km resolution in Tibet were obtained by spatial correction and downscaling of SMARTS model.On this basis,the spatial and temporal distribution characteristics of solar energy resources in the region in the past 30 years(1991–2020)are finely evaluated,and the annual global horizontal radiation resource is calculated.The results show that:1)The average annual global horizontal radiation amount in Tibet is 1816 kWh/m^(2).More than 60%of the area belongs to the“Most abundant”(GHI≥1750 kWh/m^(2))area of China’s solar energy resources category A,and nearly 40%belongs to the“Quite abundant”(1400≤GHI<1750)area of China’s solar energy resource category B.2)In space,the solar energy resources in Tibet increased gradually from north to south and from east to west.Lhasa,Central and Eastern Shigatse,Shannan,and Southwestern Ali are the most abundant cities,with a maximum annual radiation level of 2189 kWh/m2.3)In terms of time,the total horizontal radiation in Tibet was the highest in May and the lowest in December.74%of the total area belongs to the“Very stable”(R_(w)≥0.47)area of solar resource stability category A,and 26%belongs to the“stable”(0.36≤R_(w)<0.47)area of solar resource stability category B.Solar energy resources in the region show the characteristics of both strong and stable.Average solar energy resources in the region have shown a fluctuating downward trend over the past 30 years,with an average decline of about 12.86(kWh/m2)per decade.4)In terms of solar radiation resources reaching the earth’s surface,the theoretical total amount of annual horizontal radiation in Tibet is about 240.07 billion tons of standard coal or 222.91 billion kilowatts on average.

Characteristics of Desertification Change in Lake Basin Area in Gangcha County

Qinghai Lake Basin area in Gangcha county is selected as the study area in terms of desertification change features in this paper.Based on the remote sensing(RS)and global positioning system(GPS)technologies,the desertification information range from 1989 to 2014 in the study area is extracted.Using the method of the decision tree,the desertification in the research area is been divided into four grades including mild desertification,moderate desertification,severe desertification and serious desertification.The change characteristics of desertification in the study area were analyzed in detail,which showed that the desertification in the study area experienced a process of first development and then a reversal.The rapid development of desertification appears in the 1990s,where about 1101.22 kilometers of desertification area was increased in this stage.Since the twenty-first Century,the desertification is gradually significant recovered and local area exist intensified desertification.There are tendencies of interactive transform in different types of desertification.The tendencies of different degrees of desertification land are rising,and there’re some differences in rising rates,where the expansion rate of moderate desertification is the biggest,increasing by 7.27 kilometers per year.

Evolution of Desertification Types on the North Shore of Qinghai Lake

Land desertification is a widely concerned ecological environment problem.Studying the evolution trend of desertification types is of great significance to prevent and control land desertification.In this study,we applied the decision tree classification method,to study the land area and temporal and spatial change law of different types of desertification in the North Bank of Qinghai Lake area from 1987 to 2014,based on the current land use situation and TM remote sensing image data of Haiyan County,Qinghai Province,The results show that the area of mild desertification land and moderate desertification land in the study area has decreased,while the area of severe desertification land and extreme desertification land has increased significantly in the past 30 years.The area of desertification land decreased by 4.02 km2,of which the area of mild and moderate desertification land decreased by 39.73 km2 and 36.8 km2 respectively,and the area of severe and extreme desertification land increased by 32.78 km2 and 39.73 km2 respectively.As for the mutual transformation relationship,the transformation from severe desertification land to extreme desertification land is the main,and the junction of severe desertification land and extreme desertification land is the sensitive area of transformation.In the north shore of Qinghai Lake,the sandy land tends to expand eastward.The research provides reference basis for local land desertification monitoring,and has a great guidance for local effective land desertification and soil and water conservation.

Classification of Desertification on the North Bank of Qinghai Lake

In this paper,RS,GIS and GPS technologies are used to interpret the remote sensing images of the north shore of Qinghai Lake from 1987 to 2014 according to the inversion results of vegetation coverage(FVC),albedo,land surface temperature(LST),soil moisture(WET)and other major parameters after image preprocessing,such as radiometric correction,geometric correction and atmospheric correction.On this basis,the decision tree classification method based on landsat8 remote sensing image is used to classify the desertification land in this area,and the development and change of desertification in this period are analyzed.The results show that the fluctuation of desertification land area in this area increased during the study period,but from 2003 to 2014,the land area of mild desertification,moderate desertification and severe desertification landwere respectively decreased 0.92,145.89 and 29.39 km2,while the area of serious desertification land still has a slow increasing trend.Whether the driving force of desertification change trend in this area is caused by human factors or global change needs to be further studied.