Springer Science and Business Media LLC

Hydrological impacts of global climate change on regional scale are generally assessed by downscaling large-scale climatic variables, simulated by General Circulation Models (GCMs), to regional, small-scale hydrometeorological variables like precipitation, temperature, etc. In this study, we propose a new statistical downscaling model based on Recurrent Neural Network with Long Short-Term Memory which captures the spatio-temporal dependencies in local rainfall. The previous studies have used several other methods such as linear regression, quantile regression, kernel regression, beta regression, and artificial neural networks. Deep neural networks and recurrent neural networks have been shown to be highly promising in modeling complex and highly non-linear relationships between input and output variables in different domains and hence we investigated their performance in the task of statistical downscaling. We have tested this model on two datasets—one on precipitation in Mahanadi basin in India and the second on precipitation in Campbell River basin in Canada. Our autoencoder coupled long short-term memory recurrent neural network model performs the best compared to other existing methods on both the datasets with respect to temporal cross-correlation, mean squared error, and capturing the extremes.

The statistical characteristics of precipitation play an extremely important role in the risk assessment of drought and flood disasters and water resource management. In this paper, the precipitation concentration degree (PCD), precipitation concentration period (PCP) and precipitation concentration index (PCI) were used to analyse the spatiotemporal variation in precipitation concentration characteristics in China from 1960 to 2015. In addition, the cross-wavelet transform was used to analyse the possible dependencies and spatial characteristics between these three precipitation concentration indices (PCD, PCP and PCI) and monsoon indices (East Asian summer monsoon index, EASMI; South Asian summer monsoon index, SASMI; and South China Sea summer monsoon index, SCSSMI) of nine river basins in China. The results show that the spatial distribution of precipitation concentration indices in China has obvious north-south and east-west variability. China’s precipitation concentration indices are generally declining, with weak volatility. Among these indices, the PCD (P < 0.05) and PCI (P < 0.01) have decreased significantly at rates of − 0.005/10 year and − 0.006/10 year, respectively, while there has been no obvious decrease in the PCP. Compared with the PCD before 1978, the PCD in the rainy season after 1978 has decreased. In most areas, the rainy season is concentrated in July every year; however, the rainy season has been delayed in the middle and lower reaches of the Yangtze River and has advanced mainly in the other basins. The PCI has declined since 1978, indicating a decreasing contribution of heavy precipitation to the total annual precipitation, but this change in the PCI has occurred mainly at high-value stations (PCI > 0.7), and the PCI at low-value stations has increased. Most of China’s precipitation concentration indices and monsoon indices, as well as the NWRB, have small-scale (2–4 years) or medium-scale (12–15 years) oscillation periods. The oscillation periods between the precipitation concentration indices and monsoon indices in the Yangtze and Pearl River basins are more significant than those in other basins. The results of this study can help to understand the differences among the precipitation concentration characteristics in different basins in China and the intrinsic relationship between these characteristics and the summer monsoon and provide a reference for further research.

A new method is proposed to compile 1 km grid data of monthly mean air temperature by dynamically downscaling general circulation model (GCM) data with a regional climate model (RCM). The downscaling method used is a technique referred to as the pseudoglobal warming method to reduce GCM bias. For the grid data, RCM data were corrected with data from an existing meteorological network. The correction model for the RCM bias was developed by stepwise multiple regression analysis using the difference in the monthly mean air temperatures between the observation and RCM output as a dependent variable and the geographical factors as independent variables. Our method corrected the RCM bias from 1.69°C to 0.58°C for the month of August in the 1990s (1990–1999).

The analysis of trends in hydroclimatic parameters and assessment of their statistical significance have recently received a great concern to clarify whether or not there is an obvious climate change. In the current study, parametric linear regression and nonparametric Mann–Kendall tests were applied for detecting annual and seasonal trends in the relative humidity (RH) and dew point temperature (T dew) time series at ten coastal weather stations in Iran during 1966–2005. The serial structure of the data was considered, and the significant serial correlations were eliminated using the trend-free pre-whitening method. The results showed that annual RH increased by 1.03 and 0.28 %/decade at the northern and southern coastal regions of the country, respectively, while annual T dew increased by 0.29 and 0.15°C per decade at the northern and southern regions, respectively. The significant trends were frequent in the T dew series, but they were observed only at 2 out of the 50 RH series. The results showed that the difference between the results of the parametric and nonparametric tests was small, although the parametric test detected larger significant trends in the RH and T dew time series. Furthermore, the differences between the results of the trend tests were not related to the normality of the statistical distribution.

This study analyzes linear trends and change points in meteorological variables including (a) sea surface temperature (SST) anomalies using three data sets: (1) level 4 of the optimum interpolated SST (OISST) of the Advanced Very High Resolution Radiometer (AVHRR), (2) the Objectively Analyzed Air-Sea Fluxes (OAFluxSST) project, and (3) the observed Improved Extended Reconstructed SST version 4 (ERSST); (b) evaporation (Evap); (c) air temperature at 2 m (AT2m); (d) wind speed at 10 m (wind10m); (e) specific humidity at 2 m (hum2m); and (f) precipitation used from Multi-Source Weighted-Ensemble Precipitation (MSWEP) over the south of the Caspian Sea (SCS) for the period 1982–2016. A significant upward trend is observed in annual SST anomalies over the SCS at the 5% level such that OISST, OAFluxSST, and ERSST data sets, respectively, indicating warming rates of about 1.2, 1.3, and 1.9 °C during the study period. Also, a significant upward trend is observed in annual Evap, AT2m, wind10, and hum2m at the 5% level. Both the Student t parametric and Mann-Whitney nonparametric Change Point Models (CPMs) suggested occurrence of common change point for annual OISST in 1994 over the SCS. For the same year, the applied Mann-Whitney test detected change point in annual OAFluxSST. The detected change point for OISST coincides with the change points of the annual hum2m and AT2m anomalies of synoptic stations (Bandar Anzali, Babolsar, and Ramsar), located at the southern coast of the Caspian Sea. The applied CPMs suggested significant change point in Evap in 1997, as evidenced by the increase of 60 mm in the mean annual value. The annual wind10m time series have a change point in 1995. The likelihood ratio test (LRT) confirms the CPM results. For example, the LRT indicates that the annual OISST and OAFluxSST anomalies over the periods 1983–1994 and 1995–2016 are drawn from normal distributions with different means such that difference in annual OISST and OAFluxSST anomalies means before and after 1994 are 0.89 and 0.83 °C, respectively.

The study analyzed the time series of the tropical cyclone (TC) frequencies which passed through the East China Sea between July and September from 1963 to 2012. The result of applying the statistical change-point analysis to this time series shows that a climate regime shift occurred in 1983 when the TC frequencies which pass the East China Sea area started increasing. The study then analyzed the average difference after 1983 (1984–2012) and before 1983 (1963–1983). The TC genesis frequency shows a tendency in mainly appearing in the tropical and subtropical Northwestern Pacific between 1963 and 1983 and the southern part between 1984 and 2012. The TC passage frequency shows a pattern that the TCs move from the far northeast sea of Philippines and change direction to Korea and Japan, passing through the East China Sea between 1984 and 2012. Meanwhile, the TC passage frequency shows a pattern which moves from the far southeast sea of the Philippines to southern China in the west direction in the previous period (1963–1983). These TC movement patterns coincide with the development status of the subtropical western North Pacific high (SWNPH) which averages for each period. It shows that the SWNPH in the second period stays away from the SWNPH in the second period from the northeast direction, but that the SWNPH in the first period expands to western Taiwan. This study analyzes the difference between the two periods in the 500-hPa streamline to understand the changes in such TC activities in the two groups. The anomalous anticyclonic circulations centered in the southern part of Japan are fortified in most of the subtropical Northwestern Pacific. The anomalous southerlies from the anomalous circulations are outstanding in the East China Sea area, Korea, and Japan. Therefore, the TCs generated in the tropical and subtropical Northwestern Pacific move along with the anomalous steering flow (anomalous southwesterlies) and up toward the East China Sea area, Korea, and Japan. The anomalous anticyclonic circulations are fortified in most of the subtropical Northwestern Pacific areas, but the anomalous cyclonic circulations are strengthened in the tropical Northwestern Pacific below 15° N, causing the generation of TCs mainly in the northwestern part of the tropical and subtropical Northwestern Pacific between 1963 and 1983, and in the southeastern part between 1984 and 2012.

The annual variation of rainfall at 131 stations distributed all over Central America is subjected to harmonic analysis. Maps present the total variance of the mean monthly rainfall, the percentage variance of the first three harmonics, the residual contributed by the higher harmonics, the amplitudes and phase angles of the first three harmonics, and the ratio of the amplitudes of the first to te second harmonics. Annual rainfall curves are in most areas adequately reproduced by the first three harmonics. The various maps illustrate the marked regional contrasts between the Caribbean and the Pacific sides of the Isthmus and portray in particular the regional distribution of the secondary precipitation minimum around the middle of the rainy season. This quantitative description supplements an earlier study of rainfall distribution and regime in Central America.

On étudie deux modèles aléatoires stationnaires persistants parmi lesquels l'un est autorégressif du premier ordre et l'autre est à persistance finie. Tous deux sont appliqués aux valeurs diurnes de la température de l'air à Uccle et les conclusions sont favorables au schéma autorégressif. Dans ces conclusions, l'examen des caractères de dépendance de la série initiale et de ceux des séries dérivées joue un rôle prépondérant.

In the present study, we aimed to understand the current condition of land suitability and how climate change will affect its suitability for rice paddy cultivation in the Edirne Province of Turkey in the future. We used RS and the GIS-supported FAO Maximum Limitation Approach to perform land suitability analysis for the current conditions and 20-year periodic times from 2020 to 2100. The results of the current land suitability assessment indicated that 81.39% of the study area is suitable for rice paddy cultivation. Two climate change models (HadGEM2-ES and MPI-ESM-MR) and related scenarios RCP4.5 and RCP8.5 showed that the climate conditions in the region will change significantly, therefore, the suitable lands for rice paddy cultivation in the study area will increase. However, the amount of change varies across models and scenarios. Further land suitability for rice paddy cultivation in the study area will be positively affected by temperature and solar radiation changes and negatively affected by changes in humidity and precipitation. Lastly, the agricultural irrigation infrastructure is expected to be unfavorably affected by an increase in extreme climatic events. These findings can guide policymakers and stakeholders to select suitable land for future rice paddy cultivation. To adapt to climate change and reduce its effects, we recommend choosing an agricultural production model that is suitable for climate change scenarios.