Analysis of mid-twentieth century rainfall trends and variability over southwestern Uganda

A methodology has been applied to investigate the spatial variability and trends existent in a mid-twentieth century climatic time series (for the period 1943–1977) recorded by 58 climatic stations in the Albert–Victoria water management area in Uganda. Data were subjected to quality checks before further processing. In the present work, temporal trends were analyzed using Mann–Kendall and linear regression methods. Heterogeneity of monthly rainfall was investigated using the precipitation concentration index (PCI). Results revealed that 53 % of stations have positive trends where 25 % are statistically significant and 45 % of stations have negative trends with 23 % being statistically significant. Very strong trends at 99 % significance level were revealed at 12 stations. Positive trends in January, February, and November at 40 stations were observed. The highest rainfall was recorded in April, while January, June, and July had the lowest rainfall. Spatial analysis results showed that stations close to Lake Victoria recorded high amounts of rainfall. Average annual coefficient of variability was 19 %, signifying low variability. Rainfall distribution is bimodal with maximums experienced in March–April–May and September–October–November seasons of the year. Analysis also revealed that PCI values showed a moderate to seasonal rainfall distribution. Spectral analysis of the time components reveals the existence of a major period around 3, 6, and 10 years. The 6- and 10-year period is a characteristic of September–October–November, March–April–May, and annual time series.     

Spatiotemporal variations in rainfall erosivity during the period of 1960–2011 in Guangdong Province,southern China

Rainfall erosivity, which shows a potential risk of soil loss caused by water erosion, is an important factor in soil erosion process. In consideration of the critical condition of soil erosion induced by rainfall in Guangdong Province of southern China, this study analyzed the spatial and temporal variations in rainfall erosivity based on daily rainfall data observed at 25 meteorological stations during the period of 1960–2011. The methods of global spatial autocorrelation, kriging interpolation, Mann–Kendall test, and continuous wavelet transform were used. Results revealed that the annual rainfall erosivity in Guangdong Province, which spatially varied with the maximum level observed in June, was classified as high erosivity with two peaks that occur in spring and summer. In the direction of south–north, mean annual rainfall erosivity, which showed significant relationships with mean annual rainfall and latitude, gradually decreased with the high values mainly distributed in the coastal area and the low values mainly occurring in the lowlands of northwestern Guangdong. Meanwhile, a significant positive spatial autocorrelation which implied a clustered pattern was observed for annual rainfall erosivity. The spatial distribution of seasonal rainfall erosivity exhibited clustering tendencies, except spring erosivity with Moran’s I and Z values of 0.1 and 1.04, respectively. The spatial distribution of monthly rainfall erosivity presented clustered patterns in January–March and July–October as well as random patterns in the remaining months. The temporal trend of mean rainfall erosivity in Guangdong Province showed no statistically significant trend at the annual, seasonal, and monthly scales. However, at each station, 1 out of 25 stations exhibited a statistically significant trend at the annual scale; 4 stations located around the Pearl River Delta presented significant trends in summer at the seasonal scale; significant trends were observed in March (increasing trends at 3 stations), June (increasing trends at 4 stations located in the Beijiang River Basin), and October (decreasing trends at 4 stations) at the monthly scale. In accordance with the mean annual rainfall over Guangdong Province, the mean annual rainfall erosivity showed two significant periodicities of 3–6 and 10–12 years at a confidence level of 95 %. In conclusion, the results of this study provide insights into the spatiotemporal variation in rainfall erosivity in Guangdong Province and support for agrolandscape planning and water and soil conservation efforts in this region.     

Changes in daily and monthly rainfall in the Middle Yellow River,China

Highly concentrated precipitation, where a large percentage of annual precipitation occurs over a few days, may include a high risk of flooding and severe soil erosion. Thus, areas with severe erosion such as the Loess Plateau in China are particularly vulnerable to highly concentrated precipitation events due to climate change. In this study, we investigated spatial and temporal patterns in the concentration of rainfall in the Middle Yellow River (MYR) from the last 56 years (1958–2013). We used daily and monthly precipitation data from 26 meteorological stations in the study area to calculate the precipitation concentration index (PCI) and the concentration index (CI). The southern and northern parts of the MYR were characterized by a lower CI with a decreasing trend, while the middle parts had a higher CI with an increasing trend. High PCI values occurred in the southern MYR, while lower PCIs with a more homogenous rainfall distribution were found mainly in the northern parts of the MYR. The annual PCI and CI exhibited positive trends at most stations, although only a minority of stations had significant trends (P < 0.05). At seasonal scales, CI exhibited significantly increasing trends in winter at most stations, while a few stations had significant trends in the other three seasons. These findings provide important reference information to facilitate ecological restoration and farming operations in the study region.     

Rainfall and rainy days trend in Iran

In this study, long-term annual and monthly trends in rainfall amount, number of rainy days and maximum precipitation in 24?h are investigated based on the data collected at 33 synoptic stations in Iran. The statistical significance of trend and climate variability is assessed by the Mann-Kendall test. The Linear trend analysis and the Mann-Kendall test indicate that there are no significant linear trends in monthly rainfall at most of the synoptic stations in Iran. However, the maximum number of stations with negative trends have been observed in April (29 station), and then in May (21 stations) and February (21 stations) and with positive trends in December (26 stations) and July (24 stations). The significant linear trends, with a significant level of 0.05, in annual rainfall have been noticed only at five stations. The monthly number of rainy days does not show any significant linear trend for most areas in Iran. The maximum number of stations with monthly negative trends in rainy days has also been observed in April with the minimum in December. In April, out of 24 stations with negative trends, 12 stations have a significant negative trend. Contrary to that, in October there is no significant linear trend. Most stations have positive trends in annual number of rainy days. Also, the monthly maximum precipitation in 24?h does not show any significant linear trend for most areas in Iran. The maximum number of stations with monthly negative trends in maximum precipitation has also been observed in February with the minimum in December. In spite of that, there are almost no significant precipitation variations in Iran during the last 50-odd years, the tendency of decreasing rainfall amount in April and increasing rainfall amount in December and July could indicate an eventual climate change in this area in the future.     

Geographic Variation in Growing Season Rainfall during three Decades in Nigeria Using Principal Component and Cluster Analyses

Summary  Reports of changes in the seasonal and annual rainfall in Nigeria suggests that a more detailed analyses of the geographic extent of these changes and of their impact on agriculture could be of value. Variation in the growing season (April to September) rainfall from stations across Nigeria was analysed over the 30-yr period, 1960–90. Regression analyses were used to examine long-term trends. Principal component and cluster analyses were used to group stations with similar trends in standardised seasonal rainfall. Mean accumulated standardised seasonal rainfall were used to examine short- and medium-term trends for each of the groups identified. Significant (P ≤ 0.05) decreases in rainy season rainfall were found at 8 stations mostly in the Guinea and arid/semi-arid savannas of northern Nigeria, whereas no station showed significant increases. Examination of the monthly (April through September) rainfall showed that only three – Kano, Sokoto and Potiskum in the arid/semi-arid savanna – of the twenty-three stations used in the analysis had declining rainfall trends for each of the months April to September and subsequently declining seasonal rainfall trends. However, 12 to 15 stations had consistently declining rainfall trends in atleast some but not all the growing season months. However, a similar pattern was not the case in terms of increasing rainfall trends, where only one to three stations had consistently increasing rainfall trends in some but not all of the months from April to September. Stations that showed increasing rainfall trends were in the southern parts of Nigeria. Six groups with similar patterns in standardised seasonal rainfall were identified by Principal Component and Cluster analyses. For most of the groups, the period from 1967 to 1973 was that of consistently below average seasonal rainfall. However, the timing and extent of the decline varied with location. Common to stations in four of the six groups was a negative trend in seasonal rainfall for the period considered. The geographic variation in seasonal rainfall trends has tremendous agricultural significance since there are indications that the reliability of the season is decreasing from the humid forest zone with positive seasonal trends to the arid/semi-arid savanna with significant negative seasonal trends. Received June 24, 1998 Revised December 18, 1998     

Long-term rainfall characteristics in the Mzingwane catchment of south-western Zimbabwe

Rainfall characteristics during the annual rainy season are explored for the Mzingwane catchment of south-western Zimbabwe, for both historic period (1886–1906) and more recent times (1950–2015), based on available daily and monthly precipitation series. Annual and seasonal rainfall trends are determined using the modified Mann-Kendall test, magnitude of trends test and Sen’s slope estimator. Rainfall variability is quantified using the coefficient of variation (CV), precipitation concentration index (PCI) and standard precipitation index (SPI). Results suggest that contemporary mean annual rainfall may not have changed from that measured during the historic period of 1886–1906. However, the number of rainy days (≥ 1 mm) has decreased by 34%, thus suggesting much more concentrated and increased rainfall intensity. A notable shift in both the onset and cessation dates of the rainy season is recorded, particularly during the twenty-first century, which has resulted in a significantly reduced (p < 0.05) length of the rainy season. The combination of a reduced number of rainy days (≥ 1 mm) and a shortened rainy season suggests that long intra-season dry spells have become more common through time and have considerable negative consequences for agriculture and wetland ecosystem in the region. In addition, high spatio-temporal rainfall variability and seasonal PCI values indicate strong seasonality in the rainy season. Based on the SPI results, the El Niño Southern Oscillation (ENSO) strongly influences rainfall variability. The results further suggest high uncertainty in rain season characteristics, which requires effective planning for water needs.     

Iranian rainfall series analysis by means of nonparametric tests

The study of the trends and fluctuations in rainfall has received a great deal of attention, since changes in rainfall patterns may lead to floods or droughts. The objective of this study was to analyze the annual, seasonal, and monthly rainfall time series at seven rain gauge stations in the west of Iran for a 40-year period (from October 1969 to September 2009). The homogeneity of the rainfall data sets at the rain gauge stations was checked by using the cumulative deviations test. Three nonparametric tests, namely Kendall, Spearman, and Mann–Kendall, at the 95 % confidence level were used for the trend analysis and the Theil–Sen estimator was applied for determining the magnitudes of the trends. According to the homogeneity analysis, all of the rainfall series except the September series at Vasaj station were found to be homogeneous. The obtained results showed an insignificant trend in the annual and seasonal rainfall series at the majority of the considered stations. Moreover, only three significant trends were observed at the February rainfall of Aghajanbolaghi station, the November series of Vasaj station, and the March rainfall series of Khomigan station. The findings of this study on the temporal trends of rainfall can be implemented to improve the water resources strategies in the study region.     

Evaluating persistence and identifying trends and abrupt changes in monthly and annual rainfalls of a semi-arid region in Western India

In this study, 43-year (1965–2007) monthly and annual rainfall time series of ten rainfall stations in a semi-arid region of western India are analyzed by adopting three tests for testing normality and by applying autoregressive technique for exploring persistence. Gradual trends are identified by three tests, and their magnitudes are assessed by the Sen’s slope estimator. Also, abrupt changes are detected by using four tests and they are further confirmed by two tests. Box-whisker plots revealed that the rainfalls of June and September are right skewed for all the stations. The annual rainfalls of Bhinder, Dhariawad, and Gogunda stations are found considerably right skewed. The normality tests indicated that the rainfall of July does not deviate from the normal distribution at all the stations. However, the annual rainfall is found non-normal at five stations. The monthly rainfalls of June, July, and August have persistence respectively at three (Mavli, Salumber, and Sarada), two (Kherwara and Sarada), and one (Mavli) stations, whereas the annual rainfall has persistence at Girwa and Mavli stations. Significantly increasing trend is detected at Mavli in the rainfall of July and in the annual rainfall (p value?>?0.05), while the negative trend in August rainfall at Dhariawad is found significant (p value?>?0.10). This study revealed that the presence of serial correlation does not affect the performance of the Mann-Kendall test. Mean values of trend magnitudes for the rainfalls of June, July, August, and September are 0.3, 0.8, ?0.4, and 0.4 mm year^?1, respectively, and the overall mean value for the annual rainfall is 0.9 mm year^?1. It is found that the standard normal homogeneity test and the Pettitt test are biased towards the end of the series to locate a change point. Conversely, the Bayesian test has a tendency to look for a change point in the beginning of time series. Confirmed abrupt changes in the rainfall time series are found in the year 2003 (Bhinder) in June; years 1974 (Mavli) and 1989 (Dhariawad and Salumber) in July; years 1972 (Sarada), 1990 (Dhariawad), and 2003 (Mavli) in August; years 1977 (Dhariawad), 1991 (Sarada), and 2004 (Kotra) in September; and in the year 1972 (Mavli and Sarada stations) in the annual series. It is emphasized that the significantly increasing trend of rainfall may have linkages with climate change and/or variability. Finally, this study recommends use of multiple statistical tests for analyzing hydrologic time series in order to ensure reliable decisions.     

Long-term trend and variability of precipitation in Chhattisgarh State,India

Spatial and temporal precipitation variability in Chhattisgarh State in India was examined by using monthly precipitation data for 102 years (1901–2002) from 16 stations. The homogeneity of precipitation data was evaluated by the double-mass curve approach and the presence of serial correlation by lag-1 autocorrelation coefficient. Linear regression analysis, the conventional Mann–Kendall (MK) test, and Spearman’s rho were employed to identify trends and Sen’s slope to estimate the slope of trend line. The coefficient of variation (CV) was used to analyze precipitation variability. Spatial interpolation was done by a Kriging process using ArcGIS 9.3. Results of both parametric and non-parametric tests and trend tests showed that at 5 % significance level, annual precipitation exhibited a decreasing trend at all stations except Bilaspur and Dantewada. For both annual and monsoon precipitation, Sen’s test showed a decreasing trend for all stations, except Bilaspur and Dantewada. The highest percentage of variability was observed in winter precipitation (88.75 %) and minimum percentage variability in annual series (14.01 %) over the 102-year periods.     

Analysing temporal trends in the Indian Summer Monsoon and its variability at a fine spatial resolution

Temporal trends between 1951 and 2007 in annual Indian Summer Monsoon (ISM) precipitation, frequency of severe drought years and onset date of ISM were analysed on a 0.25°?×?0.25° grid cell basis across India using APHRODITE daily gridded precipitation data. Locations which experienced temporal trends of increasing or decreasing inter-annual variation in annual ISM precipitation and onset date of ISM were detected using the non-parametric Mann-Kendall test. A new method of defining local onset of ISM from daily precipitation data was developed to enable countrywide temporal trend analysis of onset date. India was characterised by a heterogeneous spatial distribution in the magnitude of inter-annual variation and location of significant temporal trends in the examined facets of ISM precipitation. A greater extent of the country experienced significant trends (p?<?0.05) of increasing inter-annual variation rather than simple increasing or decreasing trends in annual ISM precipitation and onset date of ISM. Field significance tests showed grid cells reporting significant trends were significant (p?<?0.05) at the global or field level (except trends of increasing, i.e. later, ISM onset date). This research provides finer spatial detail regarding trends and variation in annual ISM precipitation, severe drought years and onset date of ISM complementing recent studies on trends in extreme precipitation events over India to produce a comprehensive overview of recent behaviour of ISM precipitation. These findings will benefit water managers charged with managing water resources sustainably at a fine spatial scale (the watershed or basin level).     

Some Statistical Characteristics of Monthly and Annual Pluviometric Irregularity for the Spanish Mediterranean Coast

Summary  Seven series of monthly pluviometric amounts, sometimes exceeding recording periods of 100 years and compiled by the Instituto Nacional de Meteorología (Spain), are used to study the irregularity of the pluviometric regime along the Spanish Mediterranean coast and nearby Atlantic coast. First of all, three statistical functions (gamma, log-normal and a combination of Poisson and gamma distributions) and moment-ratio diagrams are used to model the monthly and annual empirical distributions of precipitation amounts, each distribution being tested by means of the Kolmogorov-Smirnov test. It is noteworthy that, whereas most of the monthly cases require the gamma distribution, the pluviometric behaviour of the summer months is well described by the Poisson-gamma distribution. Moreover, both the log-normal and the gamma distributions satisfactorily model empirical annual amounts. Consequently, rainfall amounts are not identically distributed along a year for each gauge tested. Second, temporal trends deduced for annual and seasonal amounts are computed and their statistical significance evaluated. The most notable fact is that, although some linear trends are close to 1 mm/year, their significance levels exceed the assumed threshold value and, excepting the winter season for Barcelona, they are considered non-significant from a statistical point of view. Finally, by again using monthly and annual amounts, three temporal irregularity indexes are computed for each pluviometric series, the temporal disparity of the rainfall patterns of the Mediterranean region being enhanced as a result. It addition to the temporal irregularity, a change with latitude is observed both in the parameters of the statistical distributions and the temporal irregularity indexes for the rain gauges analysed. The two most southerly rain gauges constitute a special case in comparison with the remaining stations, because they also receive the Atlantic influences due to their proximity to this ocean. Received February 25, 1999/Revised August 2, 1999     

Determination of seasonal rainfall variability, onset and cessation in semi-arid Tharaka district, Kenya

The study quantified rainfall variability for March–May (MAM) and October–December (OND) seasons in Tharaka district, Kenya. The parameters analysed were inter-annual variability of seasonal rainfall, onset and cessation using daily rainfall data in three agro-ecological zones’ stations. Percentage mean cumulative method was used to determine onset and cessation, and seasonal variability was estimated using rainfall variability indices. Although both seasons are highly variable, OND has been persistently below mean over time while MAM shows high within-season variability. Despite the near uniformity in the mean onset and cessation dates, the former is highly variable on an inter-annual scale. The two rainfall seasons are inherently dissimilar and therefore require specific cropping in agro-ecological zone LM4 and LM4-5. It is possible that farmers in IL5 are missing an opportunity by under-utilising MAM rainfall. The results should be incorporated in implications of climate variability and vulnerability assessment in semi-arid Tharaka district.     

Non-parametric trend analysis of the aridity index for three large arid and semi-arid basins in Iran

Currently, an important scientific challenge that researchers are facing is to gain a better understanding of climate change at the regional scale, which can be especially challenging in an area with low and highly variable precipitation amounts such as Iran. Trend analysis of the medium-term change using ground station observations of meteorological variables can enhance our knowledge of the dominant processes in an area and contribute to the analysis of future climate projections. Generally, studies focus on the long-term variability of temperature and precipitation and to a lesser extent on other important parameters such as moisture indices. In this study the recent 50-year trends (1955–2005) of precipitation (P), potential evapotranspiration (PET), and aridity index (AI) in monthly time scale were studied over 14 synoptic stations in three large Iran basins using the Mann–Kendall non-parametric test. Additionally, an analysis of the monthly, seasonal and annual trend of each parameter was performed. Results showed no significant trends in the monthly time series. However, PET showed significant, mostly decreasing trends, for the seasonal values, which resulted in a significant negative trend in annual PET at five stations. Significant negative trends in seasonal P values were only found at a number of stations in spring and summer and no station showed significant negative trends in annual P. Due to the varied positive and negative trends in annual P and to a lesser extent PET, almost as many stations with negative as positive trends in annual AI were found, indicating that both drying and wetting trends occurred in Iran. Overall, the northern part of the study area showed an increasing trend in annual AI which meant that the region became wetter, while the south showed decreasing trends in AI.     

A high-quality monthly total cloud amount dataset for Australia

A high-quality monthly total cloud amount dataset for 165 stations has been developed for monitoring and assessing long-term trends in cloud cover over Australia. The dataset is based on visual 9 a.m. and 3 p.m. observations of total cloud amount, with most records starting around 1957. The quality control process involved examination of historical station metadata, together with an objective statistical test comparing candidate and reference cloud series. Individual cloud series were also compared against rainfall and diurnal temperature range series from the same site, and individual cloud series from neighboring sites. Adjustments for inhomogeneities caused by relocations and changes in observers were applied, as well as adjustments for biases caused by the shift to daylight saving time in the summer months. Analysis of these data reveals that the Australian mean annual total cloud amount is characterised by high year-to-year variability and shows a weak, statistically non-significant increase over the 1957–2007 period. A more pronounced, but also non-significant, decrease from 1977 to 2007 is evident. A strong positive correlation is found between all-Australian averages of cloud amount and rainfall, while a strong negative correlation is found between mean cloud amount and diurnal temperature range. Patterns of annual and seasonal trends in cloud amount are in general agreement with rainfall changes across Australia, however the high-quality cloud network is too coarse to fully capture topographic influences. Nevertheless, the broadscale consistency between patterns of cloud and rainfall variations indicates that the new total cloud amount dataset is able to adequately describe the broadscale patterns of change over Australia. Favourable simple comparisons between surface and satellite measures of cloudiness suggest that satellites may ultimately provide the means for monitoring long-term changes in cloud over Australia. However, due to the relative shortness and homogeneity problems of the satellite record, a robust network of surface cloud observations will be required for many years to come.     

东北夏季降水异常的年代际、年际构成及成因分析

利用NCEE／NCAR再分析月平均海平面气压资料、中国160站月降水资料、英国气象局和英国大气资料中心月平均海表温度资料，使用周期分析、奇异值分解和旋转EOF分析方法，分析了东北夏季降水异常的年代际、年际构成及成因。结果表明：(1)东北夏季降水异常构成中年代际、年际变化相对均衡，在局地年际变率中，东、南部年代际变化略强于西、北部。(2)在年代际、年际尺度上，东亚夏季风增强(减弱)，则东北降水偏多(偏少)。(3)E1 Nino事件与东北区夏季风异常无直接联系．故与东北夏季降水关系不密切。     

Recent disruptions in the timing and intensity of precipitation in Calakmul,Mexico

This study addresses changes in the timing and intensity of precipitation from 1982 to 2016 from three meteorological stations around Calakmul, Mexico, a landscape balancing biodiversity conservation and smallholder agricultural production. Five methods were used to assess changes in precipitation: the Mann-Kendall test of annual and wet season trends; a fuzzy-logic approach to determine the onset of the rainy season; the Gini Index and Precipitation Concentration Index (PCI) to evaluate the temporal distribution of precipitation; Simple Precipitation Intensity Index (SDII) to evaluate precipitation intensity; and the Rainfall Anomaly Index (RAI) to identify the deficit or surplus of rainfall compared with the long-term mean. Overall, rainfall trends in Calakmul over this period indicate a slight increase, though results of the indices (Gini, SDII, PCI) all indicate that rainfall has become more intense and more unevenly distributed throughout the year. There was no significant trend in the onset date of rainfall or the RAI overall, though there were more pronounced crests and troughs from 2004 to 2016. Higher interannual variability and more pronounced rainfall anomalies, both positive and negative, suggest that rainfall in the Calakmul region has become more extreme. This research informs for management and livelihood strategies in the local region and offers insights for analyses of regional patterns of seasonal precipitation events in tropical landscapes worldwide.     

Streamflow trend analysis by considering autocorrelation structure,long-term persistence,and Hurst coefficient in a semi-arid region of Iran

Due to the substantial decrease of water resources as well as the increase in demand and climate change phenomenon, analyzing the trend of hydrological parameters is of paramount importance. In the present study, investigations were carried out to identify the trends in streamflow at 20 hydrometric stations and 11 rainfall gauging stations located in Karkheh River Basin (KRB), Iran, in monthly, seasonal, and annual time scales during the last 38 years from 1974 to 2011. This study has been conducted using two versions of Mann–Kendall tests, including (i) Mann–Kendall test by considering all the significant autocorrelation structure (MK3) and (ii) Mann–Kendall test by considering LTP and Hurst coefficient (MK4). The results indicate that the KRB streamflow trend (using both test versions) has decreased in all three time scales. There is a significant decreasing trend in 78 and 73 % of the monthly cases using the MK3 and MK4 tests, respectively, while these percentages changed to 80 and 70 % on seasonal and annual time scales, respectively. Investigation of the trend line slope using Theil–Sen’s estimator showed a negative trend in all three time scales. The use of MK4 test instead of the MK3 test has caused a decrease in the significance level of Mann–Kendall Z-statistic values. The results of the precipitation trends indicate both increasing and decreasing trends. Also, the correlation between the area average streamflow and precipitation shows a strong correlation in annual time scale in the KRB.     

呼和浩特近59年降水极端事件分析

利用呼和浩特气象站1951—2009年逐日降水量资料,以年序列的第90个百分位,建立了日降水量极端气候事件的阈值,检测了近59年来呼和浩特逐日降水量极端事件的出现频率,分析了极端事件阈值和日数及降水量的年际、年代际和季节变化,结果显示:①呼和浩特日降水极端事件的阈值小,为10.6mm;全年极端事件出现的频次11d。②降水极端事件主要出现在4-10月,且8月最多。③近59年来呼和浩特全年降水极端事件及其降水量没有显著的增减变化趋势,但而进入21世纪后,极端降水事件及其降水量的变率加大,降水强度明显减小。