Quantile regression and clustering analysis of standardized precipitation index in the Tarim River Basin,Xinjiang, China

Because drought is a very common and widespread natural disaster, it has attracted a great deal of academic interest. Based on 12-month time scale standardized precipitation indices (SPI12) calculated from precipitation data recorded between 1960 and 2015 at 22 weather stations in the Tarim River Basin (TRB), this study aims to identify the trends of SPI and drought duration, severity, and frequency at various quantiles and to perform cluster analysis of drought events in the TRB. The results indicated that (1) both precipitation and temperature at most stations in the TRB exhibited significant positive trends during 1960–2015; (2) multiple scales of SPIs changed significantly around 1986; (3) based on quantile regression analysis of temporal drought changes, the positive SPI slopes indicated less severe and less frequent droughts at lower quantiles, but clear variation was detected in the drought frequency; and (4) significantly different trends were found in drought frequency probably between severe droughts and drought frequency.     

Utility of coactive neuro-fuzzy inference system for pan evaporation modeling in comparison with multilayer perceptron

Estimation of pan evaporation (E _pan) using black-box models has received a great deal of attention in developing countries where measurements of E _pan are spatially and temporally limited. Multilayer perceptron (MLP) and coactive neuro-fuzzy inference system (CANFIS) models were used to predict daily E _pan for a semi-arid region of Iran. Six MLP and CANFIS models comprising various combinations of daily meteorological parameters were developed. The performances of the models were tested using correlation coefficient (r), root mean square error (RMSE), mean absolute error (MAE) and percentage error of estimate (PE). It was found that the MLP6 model with the Momentum learning algorithm and the Tanh activation function, which requires all input parameters, presented the most accurate E _pan predictions (r?=?0.97, RMSE?=?0.81?mm?day^?1, MAE?=?0.63?mm?day^?1 and PE?=?0.58?%). The results also showed that the most accurate E _pan predictions with a CANFIS model can be achieved with the Takagi–Sugeno–Kang (TSK) fuzzy model and the Gaussian membership function. Overall performances revealed that the MLP method was better suited than CANFIS method for modeling the E _pan process.     

Trends in evaporation of a large subtropical lake

In order to further investigate the capability of the Standardized Precipitation Index (SPI) to identify flood/drought events, monthly precipitation data from 26 precipitation stations and monthly discharge data from four hydrological stations from 1960 to 2006 in the Minjiang River basin were used to analyze the correlations between multiple time scales of the SPI and river discharge. The SPI series that had a maximum correlation with discharge was chosen to detect flood/drought events in the basin, and the results were compared to historical flood/drought events. The results indicated the following. (1) High Pearson correlations between the SPI and discharge were identified at shorter time scales (1 to 3 months), and the maximum correlation was found on the time scale of 2 months. (2) Five floods among the six largest historical flood events in the Minjiang River basin were identified with the 2-month SPI, but the SPI does have shortcomings in identifying more general floods. The SPI also identified major drought events that were consistent with historical data. This demonstrates that the 2-month SPI is an effective indicator for the identification of major flood/drought events in the Minjiang River basin.     

Hourly air temperature driven using multi-layer perceptron and radial basis function networks in arid and semi-arid regions

This study employed two artificial neural network (ANN) models, including multi-layer perceptron (MLP) and radial basis function (RBF), as data-driven methods of hourly air temperature at three meteorological stations in Fars province, Iran. MLP was optimized using the Levenberg–Marquardt (MLP_LM) training algorithm with a tangent sigmoid transfer function. Both time series (TS) and randomized (RZ) data were used for training and testing of ANNs. Daily maximum and minimum air temperatures (MM) and antecedent daily maximum and minimum air temperatures (AMM) constituted the input for ANNs. The ANN models were evaluated using the root mean square error (RMSE), the coefficient of determination (R ²) and the mean absolute error. The use of AMM led to a more accurate estimation of hourly temperature compared with the use of MM. The MLP-ANN seemed to have a higher estimation efficiency than the RBF ANN. Furthermore, the ANN testing using randomized data showed more accurate estimation. The RMSE values for MLP with RZ data using daily maximum and minimum air temperatures for testing phase were equal to 1.2°C, 1.8°C, and 1.7°C, respectively, at Arsanjan, Bajgah, and Kooshkak stations. The results of this study showed that hourly air temperature driven using ANNs (proposed models) had less error than the empirical equation.     

Tibetan Plateau summer precipitation: covariability with circulation indices

Relations between Tibetan Plateau precipitation and large-scale climate indices are studied based on the Standardized Precipitation Index (SPI) and the boreal summer season. The focus is on the decadal variability of links between the large-scale circulation and the plateau drought and wetness. Analysis of teleconnectivity of the continental northern hemisphere standardized summer precipitation reveals the Tibetan Plateau as a major SPI teleconnectivity center in south-eastern Asia connecting remote correlation patterns over Eurasia. Employing a moving window approach, changes in covariability and synchronizations between Tibetan Plateau summer SPI and climate indices are analyzed on decadal time scales. Decadal variability in the relationships between Tibetan Plateau summer SPI and the large-scale climate system is characterized by three shifts related to changes in the North Atlantic, the Indian Ocean, and the tropical Pacific. Changes in the North Atlantic variability (North Atlantic Oscillation) result in a stable level of Tibetan Plateau summer SPI variability; the response to changes in tropical Pacific variability is prominent in various indices such as Asian monsoon, Pacific/North America, and East Atlantic/Western Russia pattern.

     

Climatology of water excesses and shortages in the La Plata Basin

This study presents a multitemporal climatology of water excess and shortage during the 20th century in the La Plata Basin. The climatology is based on 0.5^o?×?0.5^o grid across the region. We transform monthly precipitation series for each point into index series at different time scales using the Standardized Precipitation Index (SPI). A month is under water excess (shortage) conditions at different time scales (i?=?6, 9, 12, and 18 months), when SPI[i](j)?>?1.5 (SPI[i](j)?<?1.5), where j is the current month. Trends in precipitation were determined using mean regional series of average values over the entire basin. A month when more than 30% of the total basin is under water excesses (shortages) is defined as an excess (shortage) critical month. From the vulnerability point of view, we analyzed the occurrence of critical months. The number of excess critical months increase with time scale of index, and almost all the critical months occurred after 1950 as a consequence of the low-frequency precipitation pattern. That means a noticeable increase in the vulnerability to extended excesses (more than 30% of the area under water excesses) after 1950, especially over the Upper Paraná and the Uruguay basins. For shortage critical months, the behavior depends on time scales. At large time scale (18 and 12 months), almost all the shortage critical months occurred in the period 1901–1950 and only at shorter time scale (9 and 6 months), some critical months appeared after 1950. That means a noteworthy decrease in the basin vulnerability to extended water shortage after 1950 and a moderate decrease in vulnerability to generalized shortage. If we analyze the frequency and mean duration of water excess and shortage events across the basin, we can appreciate that there is a tendency to relate higher frequency regions with regions with lower mean duration events, and conversely.     

Characterizing and comparing the control-run variability of eight coupled AOGCMs and of observations. Part 2: precipitation

This study examines the variability of annual-mean precipitation in eight AOGCMs and in observations using empirical orthogonal functions (EOFs). The leading mode of precipitation variability in both models and observations is centered around the low-latitude western Pacific Ocean and Indian Ocean, and is associated with the El Niño-Southern Oscillation (ENSO). The spatial pattern R ² correlations between model and observed EOF₁ range from 0.12 to 0.61. In the observations, the Southern Oscillation Index (SOI) is highly correlated (R ² = 0.82) with the amplitude of precipitation EOF₁, while model R ² correlations range from 0.17 to 0.83. If grid points near to those used to compute the standard SOI are used to compute alternative SO indices, the correlation with the amplitude of EOF₁ ranges from 0.40 to 0.90 when based on the index that maximizes the correlation. Spatial fields of the variation between local precipitation and the SOI or the North Atlantic Oscillation Index are also computed for each model and compared with the observed fields. The model fields have many important similarities with the observed fields.     

新疆塔城地区1961—2021年季节性干旱演变特征

利用塔城地区9个国家气象观测站1961—2021年逐月降水资料,基于标准化降水量指数(SPI)运用最小二乘法和Mann-Kendall检验等方法分析塔城地区干旱时空分布特征。结果表明：(1)年尺度上,塔城地区干旱发生频率为30.37%,轻旱发生最多,中旱次之；2/3站点SPI呈显著增大趋势,年站次比和干旱强度呈显著减小趋势,干旱程度有所减轻,在1987年之后塔城干旱程度整体偏轻。(2)季节尺度上,夏、秋、冬季在1980年代中期发生干旱减轻的突变,且秋季和冬季分别在2002年和1997年达到显著。(3)在影响范围方面,各季以局域性和全域性干旱为主,全域性干旱发生频率介于20～30%之间；在干旱强度方面,各季轻度干旱发生频率最高,中度以上干旱发生频率介于33～38%之间。(4)近60a塔城地区季节性干旱呈现影响范围缩小,强度减弱的变化趋势,尤其是冬季干旱站次比和干旱强度分别以-7.79%.(10a)-1^和-0.11.(10a)-1^的倾向率显著减小,干旱减轻趋势在四季中最为显著。     

Evolution of drought severity and its impact on corn in the Republic of Moldova

Droughts in Moldova were evaluated using meteorological data since 1955 and a long time series (1891?C2009). In addition, yields for corn (Zea mays L.), a crop widely grown in Moldova, were used to demonstrate drought impact. The main aim is to propose use of the S _i (S _i-a and S _i-m) drought index while discussing its potential use in studying the evolution of drought severity in Moldova. Also, a new multi-scalar drought index, the standardized precipitation?Cevapotranspiration index (SPEI), is tested for the first time in identifying drought variability in Moldova while comparing it with the commonly used standardized precipitation index (SPI). S _i-m, SPI, SPEI, and S _i-a indices show an increasing tendency toward more intensive and prolonged severely dry and extremely dry summer months. Drought frequency increased through six decades, which included long dry periods in the 1990s and 2000s. Moreover, the evolution of summer evapotranspiration recorded a positive and significant trend (+3.3?mm/year, R ²?=?0.46; p????0.05) between 1955 and 2009. A yield model based on the S _i-a agricultural index and historic corn yields explained 43% of observed variability in corn production when drought occurred in May, July, and August. Increasing severity of the 20-year drought during the critical part of the growing season is raising corn yield losses, as net losses have so far exceeded net gains.     

基于MET系统对SWCWARMS及GRAPES模式在西南地区的站点检验分析

应用国家基本观测站资料、自动站逐时降水资料,基于客观统计检验方法,针对降水(12h、24h累积雨量)、近地面要素(2m温度、10m风)和高空要素(风场、温度场、高度场),分别评估SWCWARMS模式和GRAPES模式对2015年西南地区预报能力,得到如下几点结论:(1)SWCWARMS模式降水ETS评分高于GRAPES模式,除24h小雨外SWCWARMS模式偏差值均高于GRAPES模式,两个模式在不同预报时效内对中雨、大雨、暴雨都表现一定程度的空报;(2)12h降水分段评分上,SWCWARMS模式TS评分均高于GRAPES模式,但SWCWARMS模式预报降水范围过大,随着预报时效增长空报多于GRAPES模式;SWCWARMS模式中雨和大雨空报大于其它量级降水,GRAPES模式对大暴雨漏报较多其它量级降水表现为空报;(3)两模式对高度场和温度场预报优于风场,对对流层中层预报优于中低层,SWCWARMS模式对高度场和温度场预报优于GRAPES模式,夏半年SWCWARMS模式均方根误差小于GRAPES模式;(4)两模式都表现出2m温度均方根误差在秋季增加而春季减小这一特征,SWCWARMS模式近地面要素均方根误差均小于GRAPES模式。      

Comparison of artificial neural network and multivariate linear regression methods for estimation of daily soil temperature in an arid region

Soil temperature (T _S) strongly influences a wide range of biotic and abiotic processes. As an alternative to direct measurement, indirect determination of T _S from meteorological parameters has been the focus of attention of environmental researchers. The main purpose of this study was to estimate daily T _S at six depths (5, 10, 20, 30, 50 and 100?cm) by using a multilayer perceptron (MLP) artificial neural network (ANN) model and a multivariate linear regression (MLR) method in an arid region of Iran. Mean daily meteorological parameters including air temperature (T _a), solar radiation (R _S), relative humidity (RH) and precipitation (P) were used as input data to the ANN and MLR models. The model results of the MLR model were compared to those of ANN. The accuracy of the predictions was evaluated by the correlation coefficient (r), the root mean-square error (RMSE) and the mean absolute error (MAE) between the measured and predicted T _S values. The results showed that the ANN method forecasts were superior to the corresponding values obtained by the MLR model. The regression analysis indicated that T _a, RH, R _S and P were reasonably correlated with T _S at various depths, but the most effective parameters influencing T _S at different depths were T _a and RH.     

El Nino/Southern oscillation signals in the global tropical ocean

The monthly mean sea surface temperature data of 6 areas are used to study the El Nino/Southern Oscillation signals in the global tropical ocean. These areas are in the 5oN-5oS latitude zone at 1) eastern Pacific (110o-l40oW), 2) western Atlantic (30o-50oW), 3) eastern Atlantic (10oW-10oE), 4) western Indian Ocean (30o-50oE), 5) central Indian Ocean (70o-90oE) and 6) far western Pacific (120o-140oE), and the data cover the 120-month period of December 1968 to November 1978.A power spectrum analysts shows that the characteristic time of the El Nino/Southern Oscillation (about 3-4 years) appears not only in the eastern Pacific but also in other areas of the tropics except for the western Pa-cific, where the spectrum is of white noise. The amplitude of oscillation in the eastern Pacific is about 4 times larger than the others, making the El Nino/Southern Oscillation signal the strongest in this area. According to a cross-spectrum analysis, there is no time lag between the variation in the central Indian Ocean and that in the eastern Pacific. These two areas oscillate simultaneously and comprise the main feature of the El Nino/ Southern Oscillation. Other tropical areas are related with time lags, as shown by correlation and coherence calculations.It should be noted that the sea surface temperature in the eastern Pacific oscillates in phase with that in the Indian Ocean, while the pressure oscillations in these two areas are out of phase with each other, according to the Southern Oscillation definition. It is suggested that the Southern Oscillation cannot be explained simply by the sea surface temperature anomalies.Variations in the far western equatorial Pacific do not have the time scale of the El Nino/Southern Oscilla-tion, perhaps because it is a buffer zone between the monsoon system and the trade wind system.     

Forecasting monthly precipitation in Central Chile: a self-organizing map approach using filtered sea surface temperature

Western South America is subject to considerable inter-annual variability due to El Ni?o–Southern Oscillation (ENSO) so forecasting inter-annual variations associated with ENSO would provide an opportunity to tailor management decisions more appropriately to the season. On one hand, the self-organizing maps (SOM) method is a suitable technique to explore the association between sea surface temperature and precipitation fields. On the other hand, Wavelet transform is a filtering technique, which allows the identification of relevant frequencies in signals, and also allows localization on time. Taking advantage of both methods, we present a method to forecast monthly precipitation using the SOM trained with filtered SST anomalies. The use of the SOM to forecast precipitation for Chillan showed good agreement between forecasted and measured values, with correlation coefficients (r ²) ranging from 0.72 to 0.91, making the combined use filtered SST fields and SOM a suitable tool to assist water management, for example in agricultural water management. The method can be easily tailored to be applied in other stations or to other variables.     

Multiple temporal scale variability during the twentieth century in global carbon dynamics simulated by a coupled climate–terrestrial carbon cycle model

A coupled climate–carbon cycle model composed of a process-based terrestrial carbon cycle model, Sim-CYCLE, and the CCSR/NIES/FRCGC atmospheric general circulation model was developed. We examined the multiple temporal scale functions of terrestrial ecosystem carbon dynamics induced by human activities and natural processes and evaluated their contribution to fluctuations in the global carbon budget during the twentieth century. Global annual net primary production (NPP) and heterotrophic respiration (HR) increased gradually by 6.7 and 4.7%, respectively, from the 1900s to the 1990s. The difference between NPP and HR was the net carbon uptake by natural ecosystems, which was 0.6 Pg C year^?1 in the 1980s, whereas the carbon emission induced by human land-use changes was 0.5 Pg C year^?1, largely offsetting the natural terrestrial carbon sequestration. Our results indicate that monthly to interannual variation in atmospheric CO₂ growth rate anomalies show 2- and 6-month time lags behind anomalies in temperature and the NiNO₃ index, respectively. The simulated anomaly amplitude in monthly net carbon flux from terrestrial ecosystems to the atmosphere was much larger than in the prescribed air-to-sea carbon flux. Fluctuations in the global atmospheric CO₂ time series were dominated by the activity of terrestrial vegetation. These results suggest that terrestrial ecosystems have acted as a net neutral reservoir for atmospheric CO₂ concentrations during the twentieth century on an interdecadal timescale, but as the dominant driver for atmospheric CO₂ fluctuations on a monthly to interannual timescale.     

Spatial and temporal variability of precipitation indices during 1961–2010 in Hunan Province,central south China

Based on daily precipitation records at 75 meteorological stations in Hunan Province, central south China, the spatial and temporal variability of precipitation indices is analyzed during 1961–2010. For precipitation extremes, most of precipitation indices suggest that both the amount and the intensity of extreme precipitation are increasing, especially the mean precipitation amount on a wet day, showing a significant positive trend. Meanwhile, both of the monthly rainfall heterogeneity and the contribution of the days with the greatest rainfall show an upward trend. When it comes to rainfall erosivity, most of this province is characterized by high values of annual rainfall erosivity. Although the directions of trends in annual rainfall erosivity at most stations are upward, only 6 of the 75 stations have significant trends. Furthermore, the spatial and temporal variation of dryness/wetness has been assessed by the standardized precipitation index (SPI). The principal component analysis (PCA) was applied to the SPI series computed on 24-month time scales. The results demonstrated a noticeable spatial variability with three subregions characterized by different trends: a remarkable wet tendency prevails in the central and southern areas, while the northern areas are dominated by a remarkable dry tendency.     

Study of method for synthetic precipitation data for ungauged sites using quantitative precipitation model

A method was developed to estimate a synthetic precipitation record for ungauged sites using irregular coarse observations. The proposed synthetic precipitation data were produced with ultrahigh hourly resolution on a regular 1 × 1 km grid. The proposed method was used to analyze selected real-time observational data collected in South Korea from 2010 to the end of 2014. The observed precipitation data were measured using the Automatic Weather System and Automated Synoptic Observing System. The principal objective of the proposed method was to estimate the additional effects of orography on precipitation introduced by ultrahigh- resolution (1 × 1 km) topography provided by a digital elevation model. The Global Forecast System analysis of the National Centers for Environmental Prediction was used for the upper-atmospheric conditions, necessary for estimating the orographic effects. Precipitation data from 48 of the more than 600 observation sites used in the study, which matched the grid points of the synthetic data, were not included in the synthetic data estimation. Instead, these data were used to evaluate the proposed method by direct comparison with the real observations at these sites. A bias score was investigated by comparison of the synthetic precipitation data with the observations. In this comparison, the number of Hit, False, Miss, and Correct results for 2010-2014 was 74738, 25778, 7544, and 367981, respectively. In the Hit cases, the bias score was 1.22 and the correlation coefficient was 0.74. The means of the differences between the synthetic data and the observations were 0.3, -3.9, -14.4, and -34.9 mm h^-1 and the root mean square errors (RMSEs) were 2.7, 8.3, 19.3, and 39.6 mm h^-1 for the categories of 0.5-10.0, 10.0-30.0, 30.0-50.0, and 50.0-100.0 mm h^-1, respectively. In addition, in each range, the 60% difference between the synthetic precipitation data and the observation data was -1.5 to +1.5, -5.0 to +5.0, -17.0 to +17.0, and -33.0 to +33.0 mm h^-1, respectively. Overall, the correlation coefficient of the synthetic precipitation data was > 0.7 for 43 of the 48 test stations and the RMSE was < 4 mm h^-1 at 31 stations. The results are significant at all evaluation stations at the 0.05 significance level.     

Observed spatiotemporal characteristics of drought on various time scales over the Czech Republic

This paper analyses the observed spatiotemporal characteristics of drought in the Czech Republic during the growing season (April to September) as quantified using the Standardised Precipitation Evapotranspiration Index (SPEI) on various time scales. The SPEI was calculated for various lags (1, 3, 6, 12, and 24 months) from monthly records of mean temperature and precipitation totals using a dense network of 184 climatological stations for the period 1961–2010. The characteristics of drought were analysed in terms of the temporal evolution of the SPEI, the frequency distribution and duration of drought at the country level, and for three regions delimited by station altitude. The driest and the wettest years during the growing season were identified. The frequency distribution of the SPEI values for seven drought category classes (in per cent) indicates that normal moisture conditions represent approximately 65 % of the total SPEI values for all time scales in all three regions, whereas moderate drought and moderate wet conditions are almost equally distributed around 10.5 %. Differences in extremely dry conditions (5 %) compared with extremely wet conditions (1.5 %) were observed with increasing SPEI time scales. The results of the non-parametric Mann–Kendall trend test applied to the SPEI series indicate prevailing negative trends (drought) at the majority of the stations. The percentage of stations displaying a significant negative trend for the 90, 95, 99, and 99.9 % confidence levels is approximately 40 %. An Empirical Orthogonal Functions (EOF) analysis was used to identify the principal patterns of variability of the SPEI during the growing season that accounted for the highest amount of statistical variance. The variance explained by the leading EOF range 66 to 56 %, whereas for EOF2 and EOF3, the value is between 7 and 11 % and between 4 and 7 %, respectively, for the SPEI is calculated for 1- to 24-month lags.     

Changes in ice-season characteristics of a European Arctic lake from 1964 to 2008

The long-term ice record (from 1964 to 2008) of an Arctic lake in northern Europe (Lake Kilpisj?rvi) reveals the response of lake ice to climate change at local and regional scales. Average freeze-up and ice breakup occurred on 9 November and 19 June, respectively. The freeze-up has been significantly delayed at a rate of 2.3 d per decade from 1964 onward (P?<?0.05). No significant change has taken place in ice breakup. Annual average ice thickness has become smaller since the mid-1980s (P?<?0.05). Air temperature during the early ice season significantly affected the ice thickness. The freeze-up date exhibits the highest correlation with the 2-month average daily minimum air temperature centered at the end of October, while the ice breakup date exhibits the highest correlation with the 2-month average daily maximal air temperature centered in mid May. A 1°C increase in the surface air temperature corresponds to a freeze-up later by 3.4?days and an ice breakup earlier by 3.6?days. Snow cover is a critical factor in lake-ice climatology. For cumulative November to March precipitation of less than 0.13?m, the insulating effect of the snow dominated, while higher rates of precipitation favored thicker ice due to the formation of snow ice. Variations in ice records of Lake Kilpisj?rvi can serve as an indicator of climate variations across the northern Europe. The North Atlantic Oscillation (NAO) does not significantly affect the ice season there, although both the local air temperatures and winter precipitation contain a strong NAO signal.     

Climate controls on rainfall isotopes and their effects on cave drip water and speleothem growth: the case of Molinos cave (Teruel,NE Spain)

The interpretation of stable isotopes in speleothems in terms of past temperature variability or precipitation rates requires a comprehensive understanding of the climatic factors and processes that influence the δ¹⁸O signal in the way through the atmosphere to the cave, where carbonate precipitates acquiring its final isotopic composition. This study presents for the first time in the Iberia Peninsula an integrated analysis of the isotopic composition of rainfall (δ¹⁸O_p) during 2010–2012 years and, through a detailed monitoring survey, the transference of the primary isotopic signal throughout the soil and epikarst into the Molinos cave (Teruel, NE Spain). Both air temperature and amount of precipitation have an important effect on δ¹⁸O_p values, clearly imprinting a seasonal variability modulated by an amount effect when rainfall events are more frequent or intense. Air mass history and atmospheric circulation influences are considered through the study of weather types, synoptic-scale climate patterns and large-scale atmospheric circulation indexes (North Atlantic Oscillation and Western Mediterranean Oscillation) revealing a dominant source effect on δ¹⁸O_p values in this region where tropical North Atlantic and Western Mediterranean are the two moisture source regions. A delay of 2–3 months occurs between the dripwater oxygen isotopic composition (δ¹⁸O_d) respect to δ¹⁸O_p values as a consequence of large residence time in the epikarst. Limited calcite precipitates are found from winter to spring when δ¹⁸O_d values are less negative and dripwater rates are constant. This study suggests that NE Iberian δ¹⁸O_calcite proxy records are best interpreted as reflecting a combination of parameters, not just paleotemperature or paleorainfall and, if extending present-day situation towards the recent past, a biased signal towards winter values should be expected in Molinos speleothem records.     

Combining the standardized precipitation index and climatic water deficit in characterizing droughts: a case study in Romania

This paper characterizes droughts in Romania using the approach of both the standardized precipitation index (SPI) and climatic water deficit (WD). The values of the main climatic factors (rainfall, temperature, reference evapotranspiration, etc.) were obtained from 192 weather stations in various regions of Romania. Penman–Monteith reference evapotranspiration (ET_o-PM) was used to calculate WD as the difference between precipitation (P) and ET_o-PM. SPI was calculated from precipitation values. There is a clear difference between drought and aridity. Drought occurrence determines higher WD values for plains and plateaus and lower climatic excess water (EW) values for high mountains in Romania, depending on the aridity of the specific region considered and drought severity. WD calculated as mean values for both normal conditions and, for all locations studied, various types of drought was correlated with mean annual precipitation and temperature, respectively. The combined approach of WD and SPI was mainly carried out for periods of 1 year, but such studies could also be done for shorter periods like months, quarters, or growing season. The most arid regions did not necessarily coincide with areas of the most severe drought, as there were no correlations between WD and SPI and no altitude-based SPI zones around the Carpathian Mountains, as is the case for other climate characteristics, soils and vegetation. Water resource problems arise where both SPI values characterize extremely droughty periods and WD values are greatly below ?200 mm/year. This combined use of SPI and WD characterizes the dryness of a region better than one factor alone and should be used for better management of water in agriculture in Romania and also other countries with similar climate characteristics.