Total Seasonal Rainfall Research Articles

Trend analysis and change point detection in precipitation time series over the Eastern Province of Rwanda during 1981–2021

This study examines trends and change points in agroclimatic variables at 56 meteorological stations’ locations and region levels in Rwanda’s Eastern Province from 1981 to 2021. We used the Mann–Kendall and Regional Kendall tests, along with Sen’s Slope and Sequential Mann–Kendall Rank Statistic tests, to analyse six key agricultural indicators: seasonal rainfall totals, number of rainy days, rainfall intensity (light, moderate, heavy), onset and cessation dates, and season duration. In the March to May (MAM) season, 39 out of 56 stations recorded a decreasing rainfall trend, with significant trends observed at eight stations in the south. Conversely, 17 stations showed increasing trends, with only one in the north being significant. Regionally, the trend was a non-significant decrease. In the September to December (SOND) season, 31 stations (one significant) experienced decreasing rainfall trends. Among the 25 stations showing increasing trends, only one was significant. The regional trend indicates a non-significant increase. Onset days showed a decreasing trend at 41 stations (12 significant) in both MAM and SOND and a significant regional trend in SOND. Season duration increased at 43 stations in MAM (five significant) and 48 stations in SOND (six significant), with the regional trend being significant only in SOND. Heavy rainy days indicate a significantly regional decreasing trend in MAM. The change point of most stations with decreasing and increasing trends occurred between 2000–2020 and 1980–2000, respectively. These fluctuations have affected agricultural practices and led to crop failures, emphasizing the region’s need for better climate information services and adaptation strategies.

Locally Defined Seasonal Rainfall Characteristics within the Horn of Africa Drylands from Rain Gauge Observations

Abstract Seasonal rainfall is critical to lives and livelihoods within the Horn of Africa drylands (HAD), but it is highly variable in space and time. The main HAD rainfall seasons are typically defined as March–May (MAM) and October–December (OND). However, these 3-month periods are only generalized definitions of seasonality across the HAD, and local experience of rainfall may depart from these substantially. Here, we use daily rain gauge data with a duration of at least 10 years from 69 stations across the drylands of Kenya, Somalia, and Ethiopia to locally delineate key rainfall seasons. By calculating local seasonal rainfall timings, totals, and extremes, we obtain more accurate estimates of the spatial variability in rainfall delivery across the HAD, as well as climatological patterns. Results show high spatial variability in season onset, cessation, and length across the region, indicating that a homogenous classification of rainfall seasons across the HAD (e.g., MAM and OND) is inadequate for representing local rainfall characteristics. Our results show that the “long rains” season is not significantly longer than the “short rains” season over the period of study. This could be related to the previously documented decline of the “long rains” seasonal totals over recent decades. Several rainfall metrics also vary spatially between seasons, and the rainfall on the most extreme days can accumulate to double the local mean seasonal total. The locally defined rainfall seasons better capture the bulk of the rainfall during the season, giving improved characterization of rainfall metrics, consistent with the aim of a better understanding of rainfall impacts on local communities.

How unusual was Australia's 2017–2019 Tinderbox Drought?

Australia's Murray-Darling Basin experienced three consecutive years of meteorological drought across 2017–2019, collectively named the ‘Tinderbox Drought’. Rainfall deficits during the three-year drought were most pronounced in the Australian cool season (April to September). Deficits in both the cool season and annual total rainfall were unprecedented in the instrumental record. However, the instrumental record provides just one of a range of equally plausible climate trajectories that could have occurred during this period. To determine if the Tinderbox Drought was outside this range, we used observational data from prior to the onset of the drought to construct Linear Inverse Models (LIMs) that emulate the stationary statistics of Australian rainfall and its connection to global sea surface temperature (SST) anomalies. Overall, we find that rainfall deficits were most unusual in the northern Murray-Darling Basin, and during the final year of the drought. The global SST anomalies observed during the first two years of the Tinderbox Drought, particularly the cool anomalies in the central tropical Pacific and western Indian Ocean, are not typically associated with low rainfall across the Murray-Darling Basin. In terms of single-year rainfall anomalies, the only aspect of the Tinderbox Drought that was beyond the range of the LIMs was annual-total rainfall over the northern Murray-Darling Basin during 2019. This coincided with an extreme positive Indian Ocean Dipole event that was also beyond the range of the LIMs. When considered in terms of basin-wide rainfall over the full three years, rainfall deficits during the Tinderbox Drought were beyond the LIM range in terms of both cool-season and annual-total rainfall. This suggests an anthropogenic contribution to the severity of the drought—likely exacerbated by the 2019 extreme positive Indian Ocean Dipole event.

Using the Observed Variations of the Start Date of the Rainy Season over Central America for Its Reliable Seasonal Outlook

Abstract We introduce a simple method to define the start and the end of the rainiest part of the year as the first and the last day of the year when the daily rain rate is more or less than the annual mean climatological rain rate for a region or at a given grid point of the rainfall analysis, respectively. A novelty of this work is the adoption of a perturbation technique to generate a total of 1001 ensemble members to account for observational and analysis uncertainties. This allows for a probabilistic estimate of the start and retreat dates of the rainy season at the granularity of the Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (IMERG), version 6, rainfall analysis over Central America. The seasonal cycle of the IMERG rainfall analysis is also found to verify with in situ observations in the region. Many large-scale climate drivers affect regional rainfall, often with complex interactions that affect the onset date, retreat date, and magnitude of the seasonal rainfall cycle, making it difficult to predict the length or total quantity of seasonal rainfall using climate drivers alone. Once an onset date is established, however, this metric alone can be more indicative of both the length and total seasonal rainfall anomaly than predicting how the climate drivers will interact to affect the quantity and duration of upcoming seasonal rainfall. The local relationships of the start date with seasonal length and rainfall anomaly are leveraged to produce effective seasonal outlooks of the rainy season for the region by just monitoring the start date variations. Significance Statement The start date and retreat date of the rainy season in Central America are defined every year, precisely to a specific date from our proposed definition of it. This is possible because the region exhibits a strong seasonality of the rainfall. As a result, the year-to-year (interannual) variation of the seasonal rainfall during the rainy season is also determined by the variations in the length of the season that are usually overlooked in fixed calendar seasons. We define the start or retreat date of the rainy season as the first or the last day of the year when the daily rain rate exceeds or falls below the average daily rainfall from 2001 to 2022, respectively. We also find that both start and retreat date variations independently influence the length and total seasonal rainfall of the rainy season. Consequently, these relationships are leveraged to provide an outlook of the forthcoming rainy season from the diagnosis of the variations of its onset date, which is shown to be an effective predictor with significant useful seasonal prediction skills.

Variability of rainy season onsets over East Africa

AbstractOver the East Africa region forecasts of the onset of the rainy seasons have the potential to support decision‐making, especially in the largely rain‐fed agricultural sector. However, the understanding of key features of onset remains limited. Here, we analyse the variability of onset and associated drivers at interannual and subseasonal timescales, using several onset definitions. Results show that the onset date is especially variable from year to year in some of the high‐potential agricultural areas (standard deviation >20 days), which has implications for agricultural risk management. The choice of onset definition metric matters; agronomic definitions have limited applicability at the regional scale and are also highly sensitive to the spatial scale of analysis and to the choice of rainfall data. Onset information provided at coarse scales should be used with caution for decision‐making at the local scale; the “hit rate” of coarse‐scale tercile onset information at the local scale is less than 40% on average. To varying degrees, onset is related to total seasonal rainfall and thus to dominant interannual drivers of rainfall, including the Indian Ocean Dipole and ENSO modes in October–December and the western Pacific “V‐gradient” pattern in March–May. However, by analysing the dominant proportion of onset variance unrelated to total rainfall during the climatological season we show a substantial influence of subseasonal drivers, notably the Madden–Julian Oscillation. As such, there is an opportunity for rainfall onset information to be provided across seasonal and subseasonal timescales. Our work reinforces the need for enhanced co‐production of such onset information with stakeholders, especially regarding the choice of metric, alignment of forecasts with livelihood calendars, interpretation of the credibility of information content for local‐level decision‐making, as well as appropriate strategies for staggered risk management interventions informed by forecasts over “seamless” lead times.

Reassessing and Extending the Composite Rainfall Record of Manchester, Northwest England: 1786–Present

A monthly composite rainfall record for the period 1786–present representative of Manchester, northwest England is presented. The 235-year record ranks as the second-longest instrumental rainfall record available for northern England, and the fifth-longest for the UK, and contributes to a growing network of long homogenous rainfall series. A composite record is constructed, extended, and homogenised, and the record is analysed in terms of annual and seasonal variability, with a focus on extreme wet/dry events. Three primary meteorological stations in Manchester, located within 2 km of one another, form the basis of the reconstruction, with other records identified for infilling and extension based on their longevity, continuity, and proximity to the primary stations. A linear regression analysis is applied to produce a continuous record, and adjustment factors are applied to ensure homogeneity. Record homogeneity is assessed via cross-comparison with long-term records from the region (Carlisle, Chatsworth House and HadNWEP), and the methods are applied to assess relative homogeneity include the double-mass curve and Standard Normal Homogeneity tests. The Manchester record is deemed to be homogenous overall but includes two periods of increased uncertainty: 1786–1819, comprising the earliest observations and greatest number of different stations, and 1883–1911, which encompasses multi-year and multi-decadal drought events of (1883–1885 and 1890–1910) as identified by other long-term meteorological studies. The analysis of the entire record reflects long-term rainfall variability with an increasing, although not significant, trend in annual rainfall observed. Seasonally, a significant increase in winter rainfall is exhibited, in keeping with patterns observed in other regional studies. Seasonal rainfall totals are found to be highly variable at the decadal timescale. Several well-documented extreme wet (e.g., autumn 2000) and dry (e.g., summer 1976) seasons are identified, including historic events (e.g., the floods of summer 1872 and drought of summer 1887) from the less-well documented eighteenth and nineteenth centuries.

Impact of Atlantic multidecadal variability on rainfall intensity distribution and timing of the West African monsoon

Abstract. Previous studies agree on an impact of the Atlantic multidecadal variability (AMV) on the total seasonal rainfall amounts over the Sahel. However, whether and how the AMV affects the distribution of rainfall or the timing of the West African monsoon is not well known. Here we seek to explore these impacts by analyzing daily rainfall outputs from climate model simulations with an idealized AMV forcing imposed in the North Atlantic, which is representative of the observed one. The setup follows a protocol largely consistent with the one proposed by the Component C of the Decadal Climate Prediction Project (DCPP-C). We start by evaluating model's performance in simulating precipitation, showing that models underestimate it over the Sahel, where the mean intensity is consistently smaller than observations. Conversely, models overestimate precipitation over the Guinea coast, where too many rainy days are simulated. In addition, most models underestimate the average length of the rainy season over the Sahel; some are due to a monsoon onset that is too late and others due to a cessation that is too early. In response to a persistent positive AMV pattern, models show an enhancement in total summer rainfall over continental West Africa, including the Sahel. Under a positive AMV phase, the number of wet days and the intensity of daily rainfall events are also enhanced over the Sahel. The former explains most of the changes in seasonal rainfall in the northern fringe, while the latter is more relevant in the southern region, where higher rainfall anomalies occur. This dominance is connected to the changes in the number of days per type of event; the frequency of both moderate and heavy events increases over the Sahel's northern fringe. Conversely, over the southern limit, it is mostly the frequency of heavy events which is enhanced, thus affecting the mean rainfall intensity there. Extreme rainfall events are also enhanced over the whole Sahel in response to a positive phase of the AMV. Over the Sahel, models with stronger negative biases in rainfall amounts compared to observations show weaker changes in response to AMV, suggesting that systematic biases could affect the simulated responses. The monsoon onset over the Sahel shows no clear response to AMV, while the demise tends to be delayed, and the overall length of the monsoon season enhanced between 2 and 5 d with the positive AMV pattern. The effect of AMV on the seasonality of the monsoon is more consistent to the west of 10∘ W, with all models showing a statistically significant earlier onset, later demise, and enhanced monsoon season with the positive phase of the AMV. Our results suggest a potential for the decadal prediction of changes in the intraseasonal characteristics of rainfall over the Sahel, including the occurrence of extreme events.

Decline in small mammal species richness in coastal-central California, 1997-2013.

The richness and composition of a small mammal community inhabiting semiarid California oak woodland may be changing in response to climate change, but we know little about the causes or consequence of these changes. We applied a capture-mark-recapture model to 17 years (1997-2013) of live trapping data to estimate species-specificabundances. The big-eared woodrat was the most frequently captured species in the area, contributing 58% of total captures. All small mammal populations exhibited seasonal fluctuations, whereas those of the California mouse, brush mouse, and pinyon mouse declined during the study period. We also applied a multispecies dynamic occupancy model to our small mammal detection history data to estimate species richness, occupancy (), detection (p), local extinction (), and colonization () probabilities, and to discern factors affecting these parameters. We found that decreased from 0.369 ± 0.088 in 1997 to 0.248 ± 0.054 in 2013; was lower during the dry season (May-September) than the wet season (October-April) and was positively influenced by total seasonal rainfall (slope parameter, = 0.859 ± 0.371; 95% CI = 0.132-1.587). Mean mammalian species richness decreased from 11.943 ± 0.461 in 1997 to 7.185 ± 0.425 in 2013. With highly variable climatic patterns expected in the future, especially increased frequency and intensity of droughts,itis important to monitor small mammal communities inhabiting threatened California oak woodlands.

Temporal Trend Analysis of Annual, Maximum, and Seasonal Rainfall in Selected Weather Stations in Region IV-A, Philippines

Temporal variability analysis of rainfall is important for planning agricultural crop production activities, designing water resource systems, and conducting climatological research, particularly climate change analysis. This study aimed to analyze the temporal trends of annual (total and 1-d maximum) and seasonal (total) rainfall observed in Region IV-A. Historical rainfall data were collected from seven weather stations scattered all over the region. Each dataset was initially subjected to a series of homogeneity tests (standard normal homogeneity, Pettitt’s, Buishand’s range, and von Neumann ratio), followed by trend analysis using both linear regression analysis and non-parametric methods [Mann-Kendall (MK), modified MK (MMK), and Sen’s slope (SS) estimator]. All data series proved to be homogeneous except for the annual and seasonal (DJF) total rainfall in Sangley Point station. The results of the homogeneity test in Sangley Point in annual and DJF timescale suggest the consideration of the breakpoint on the years of 1984 and 1999, respectively, for trend analysis. Temporal trend based on linear regression showed a weak correlation between rainfall of various time scales and time in years, with only three stations showing a significant trend at a 5% significance level – namely, Los Baños (DJF, 2.86 mm/yr), Tayabas (DJF, 7.99mm/yr), and Sangley Point (JJA, 9.32 mm/yr). On the other hand, both MK and SS indicates a significant increase in the seasonal (DJF) total rainfall in both stations of Ambulong (1.04 mm/yr) and Los Baños (2.68 mm/yr). Overall, the results of trend analysis revealed an increasing trend in the cumulative rainfall across timescales in most stations in the region which has an implication on the timing of the optimum cropping patterns in agricultural areas and on the hydrology of the watersheds in the region. This study has generated new knowledge and information that could prove useful for agricultural production management and for proper planning, design, and management of water resource systems.

California margin temperatures modulate regional circulation and extreme summer precipitation in the desert Southwest

In August 2022, Death Valley, the driest place in North America, experienced record flooding from summertime rainfall associated with the North American monsoon (NAM). Given the socioeconomic cost of these type of events, there is a dire need to understand their drivers and future statistics. Existing theory predicts that increases in the intensity of precipitation is a robust response to anthropogenic warming. Paleoclimatic evidence suggests that northeast Pacific (NEP) sea surface temperature (SST) variability could further intensify summertime NAM rainfall over the desert southwest. Drawing on this paleoclimatic evidence, we use historical observations and reanalyzes to test the hypothesis that warm SSTs on the southern California margin are linked to more frequent extreme precipitation events in the NAM domain. We find that summers with above-average coastal SSTs are more favorable to moist convection in the northern edge of the NAM domain (southern California, Arizona, New Mexico, and the southern Great Basin). This is because warmer SSTs drive circulation changes that increase moisture flux into the desert southwest, driving more frequent precipitation extremes and increases in seasonal rainfall totals. These results, which are robust across observational products, establish a linkage between marine and terrestrial extremes, since summers with anomalously warm SSTs on the California margin have been linked to seasonal or multi-year NEP marine heatwaves. However, current generation earth system models (ESMs) struggle to reproduce the observed relationship between coastal SSTs and NAM precipitation. Across models, there is a strong negative relationship between the magnitude of an ESM’s warm SST bias on the California margin and its skill at reproducing the correlation with desert southwest rainfall. Given persistent NEP SST biases in ESMs, our results suggest that efforts to improve representation of climatological SSTs are crucial for accurately predicting future changes in hydroclimate extremes in the desert southwest.

Calibrated multi‐model ensemble seasonal prediction of Bangladesh summer monsoon rainfall

AbstractBangladesh summer monsoon rainfall (BSMR), typically from June through September (JJAS), represents the main source of water for multiple sectors. However, its high spatial and interannual variability makes the seasonal prediction of BSMR crucial for building resilience to natural disasters and for food security in a climate‐risk‐prone country. This study describes the development and implementation of an objective system for the seasonal forecasting of BSMR, recently adopted by the Bangladesh Meteorological Department (BMD). The approach is based on the use of a calibrated multi‐model ensemble (CMME) of seven state‐of‐the‐art general circulation models (GCMs) from the North American Multi‐Model Ensemble project. The lead‐1 (initial conditions of May for forecasting JJAS total rainfall) hindcasts (spanning 1982–2010) and forecasts (spanning 2011–2018) of seasonal total rainfall for the JJAS season from these seven GCMs were used. A canonical correlation analysis (CCA) regression is used to calibrate the raw GCMs outputs against observations, which are then combined with equal weight to generate final CMME predictions. Results show, compared to individual calibrated GCMs and uncalibrated MME, that the CCA‐based calibration generates significant improvements over individual raw GCM in terms of the magnitude of systematic errors, Spearman's correlation coefficients, and generalised discrimination scores over most of Bangladesh areas, especially in the northern part of the country. Since October 2019, the BMD has been issuing real‐time seasonal rainfall forecasts using this new forecast system.

Factors affecting interannual variation in late summer rainfall in the Red River Delta of Vietnam

Based on station monthly observational statistics, the major rainfall in the Red River Delta (RRD) of Vietnam occurs in late summer (July–September) with conspicuous year-to-year variation. Using the ± 0.8 of the seasonal total rainfall standard deviation as criteria, seven wet and six dry years are identified over the period 1983–2015. In addition to the 70.5% of the seasonal total rainfall contributed by the heavy rainfall days, the distinct heavy rainfall accumulation difference between wet and dry years seems to fundamentally establish these two separated extreme wet and dry groups. As revealed from further analyses, the large variability in rainfall is attributed to the influence of tropical cyclones (TCs) and 7–24- and 30–60-day intraseasonal oscillations (ISOs); in particular, the number of TCs affecting the RRD and rainfall produced by TCs are more (less) during the wet (dry) years, and the amplitudes of ISOs are also enhanced (reduced). In many cases, heavy rainfall days are induced by the combined effect of both ISOs and TCs, while some heavy rainfall events are mainly triggered by ISOs. It is found from the water vapor budget analyses that an anomalous cyclone (anticyclone) dominates over the Indochina Peninsula in wet (dry) years, resulting in more (less) water vapor being transported to the RRD, whereas the anomalous convergence (divergence) of water vapor flux leads to the maintenance of excessive (insufficient) rainfall over the RRD. However, the El Niño–Southern Oscillation (ENSO) forcing shows minor effects on the interannual variation in rainfall in the RRD.

Changing Trends in Temperatures and Rainfalls in the Western Pacific: Guam

Pacific islands have always been at the front of the great challenge of climate change. In this study, Mann–Kendall’s tau-based slope estimator was implemented to detect statistical trends in daily maximum and minimum temperatures of 2 stations and daily rainfalls at 14 stations over Guam for the period of 1953–2021, respectively, with 17 climate change detection indices. Mann–Kendall tests were implemented to the detection indices with respect to different time frames (i.e., annual, two-seasonal, and four-seasonal). The p-values from Mann–Kendall tests were used to determine the strength of trends, and Sen’s slopes were applied for the magnitudes of trends. The temperature trend analysis results indicate that Guam’s climate is getting warmer year by year. The increasing magnitudes of a seasonal maximum of daily maximum temperatures during the dry season are 0.036 °C/year for the dry season and 0.025 °C/year for the wet season at Anderson Airforce Base, while 0.031 °C/year and 0.023 °C/year for the dry and wet seasons at Guam International Airport. Trend analyses for temperatures have indicated that temperature during April through June has been increasing rapidly compared to other seasons. Strong trends in seasonal total rainfall amounts and the number of wet days were observed from July through December. The increasing trends in extreme rainfall indices during January–March and July–September periods would aggravate water quality due to the more sediments since important ecological reserve areas and coral reef areas are linked to watersheds in southern Guam.

Characteristics of tropical–extratropical cloud bands over tropical and subtropical South America simulated by BAM‐1.2 and HadGEM3‐GC3.1

AbstractTropical–extratropical cloud bands are common in South America (SAm), contributing significantly to the total rainy season precipitation. Thus, it is fundamental that climate and weather forecast models correctly represent them and their associated dynamic aspects. Adopting an event‐based framework, we evaluate the performance of two global models in simulating the observed cloud bands over SAm: the Brazilian Global Atmospheric Model version 1.2 (BAM‐1.2) and the Hadley Centre Global Environment Model in the Global Coupled configuration 3.1 (HadGEM3‐GC3.1). Both models reproduce the main characteristics of cloud bands and the dynamical aspects leading to their development and persistence. Nonetheless, the biases in precipitation during simulated cloud bands contribute more than 50% of the bias in total precipitation in some regions. BAM‐1.2 simulates fewer but more persistent cloud bands than observed; HadGEM3‐GC3.1 simulates weaker cloud band activity during early summer and more persistent events after January than observed. In all models, the biases in cloud band events arise from the interaction between biases in the basic state and the synoptic‐scale regional circulation. In the basic state, stronger upper level westerlies over the midlatitude South Pacific support the propagation of longer and slower Rossby waves towards subtropical SAm, increasing the duration of the cloud band events. This bias interacts with negative biases in the upper level westerlies over subtropical SAm, increasing the wind shear, hindering the propagation of synoptic‐scale Rossby waves into lower latitudes, and resulting in biases in the cloud band location, intensity, and seasonality. The application in this study of an event‐based framework robust to differences in model resolution and complexity enables the identification of small but critical biases in circulation. These biases are linked to synoptic‐scale rainfall system biases and help to explain the season total rainfall model biases.

Evaluation of downscaling seasonal climate forecasts for crop yield forecasting in Zimbabwe

Meteorology and weather forecasting are crucial for water-limited agriculture. We evaluate the added value of downscaling seven-months global deterministic seasonal forecasts from the Climate Forecast System version 2 (CFSv2) using the Weather, Research and Forecasting (WRF) model over Zimbabwe for ten growing seasons (2011–2021). Downscaling reduces the area of significant differences between forecasted and observed total seasonal rainfall. Downscaling also improves the score for droughts as measured through the standardized precipitation index and 3-class method. Yield forecasts by the WOrld FOod STudies (WOFOST) model reveal that downscaling improves the estimated growing season evolution and maize yield in all studied regions across the country. For the main maize production region Karoi, the bias, root mean square error, mean absolute error and mean absolute percentage error reduce by 33% (0.2 ton/ha), 27% (0.4 ton/ha), 31% (0.4 ton/ha) and 27% (8.3 %) respectively by downscaling. Hence we illustrate that downscaling the deterministic seasonal forecasts may assist in food security in a crucial area in southern Africa.

A Machine Learning Approach for Probabilistic Multi-Model Ensemble Predictions of Indian Summer Monsoon Rainfall

Due to the uncertainty associated with Indian summer monsoon rainfall (ISMR), probabilistic seasonal forecasts which can convey the inherent uncertainty of ISMR are more useful to the user community than a single deterministic forecast. While such probabilistic seasonal forecasts can be produced from general circulation model (GCM) output, one single model generally does not represent all sources of error. The probabilistic multi model ensemble (PMME) is a well-accepted way to improve on the skill of probabilistic forecasts by individual GCMs. PMME can be constructed with one of two approaches: non-parametric, or parametric with respect to the occurrence of three categories of seasonal total rainfall—below, near, and above normal as defined by the climatological base period. Both the methods have their limitations. Non-parametric PMME use a smaller ensemble size which results in overconfident forecasts, and parametric PMME make the inaccurate assumption that total rainfall follows a Gaussian distribution. To avoid these problems, we propose the use of the Extreme Learning Machine (ELM), a novel machine learning (ML) approach, to construct PMME for ISMR forecasting. ELM is a state-of-the-art generalized form of single-hidden-layer feed-forward neural network. However, since the traditional ELM network only produces a deterministic outcome, we use a modified version of ELM called Probabilistic Output Extreme Learning Machine (PO-ELM). PO-ELM uses sigmoid additive neurons and slightly different linear programming to make probabilistic predictions. The performance of such PO-ELM based PMME is assessed rigorously in terms of Generalized Receiver Operating Characteristic scores and reliability diagrams over a 37 years period spanning from 1982 to 2018 following a leave-three-year-out cross-validation scheme. It is demonstrated that our new strategy for PMME based on ML is capable of producing skillful MME forecasts over large regions of India.

Maize Yield Response, Root Distribution and Soil Desiccation Crack Features as Affected by Row Spacing.

Plant density is among the most critical factors affecting plant yields and resource use efficiency since it drives the exploitation of the available resources per unit area, root distribution and soil water losses by direct evaporation from the soil. Consequently, in fine-textured soils, it can also affect the formation and development of desiccation cracks. The aim of this study, carried out on a sandy clay loam soil in a typical Mediterranean environment, was to investigate the effects of different row spacings of maize (Zea mais L.) on yield response, root distribution and the main features of desiccation cracks. The field experiment compared bare soil and soil cropped with maize using three plant densities (6, 4 and 3 plants m-2), obtained by keeping the number of plants in a row constant and varying the distance between the rows (0.5-0.75-1.0 m). The highest kernel yield (16.57 Mg ha-1) was obtained with the greatest planting density (6 plants m-2) with a row spacing of 0.5 m; significantly lower yields were recorded with spacings of 0.75 and 1 m, with a decrease of 8.09% and 18.24%, respectively. At the end of the growing season, soil moisture in the bare soil was on average 4% greater in comparison to the cropped soil and was also affected by row spacing, decreasing with the decrease in the inter-row distance. An inverse behaviour was observed between soil moisture and both root density and desiccation crack size. Root density decreased to the increase in soil depth and to the increase in distance from the row. The pluviometric regime occurred during the growing season (total rainfall of 343 mm)-resulted in the formation of cracks of reduced size and with an isotropic behaviour in the bare soil, whereas in the cultivated soil, the cracks were parallel to the maize rows and increased in size with decreasing inter-row distance. The total volume of the soil cracks reached a value of 135.65 m3 ha-1 in the soil cropped with a row distance of 0.5 m, and was about ten times greater in comparison to the bare soil and three times greater in comparison to a row spacing of 1 m. Such a volume would allow a recharge of 14 mm in the case of intense rainy events on soil characterised by low permeability.

Signals of northward propagating monsoon intraseasonal oscillations (MISOs) in the RegCM4.7 CORDEX-CORE simulation over South Asia domain

Northward propagating summer monsoon intraseasonal oscillations (MISOs) in the Indian Ocean region remain poorly understood and difficult to predict. Here we examine a free-running high-resolution regional atmospheric model (RegCM4.7 with 25 km resolution), forced distantly at the boundaries by atmospheric observations (ERA-Interim, 0.75 $$^{\circ }$$ ) and forced locally by observed sea-surface temperature (SST) over the period 1979–2016, to assess its ability to reproduce key aspects of these MISOs. We find that the model MISO exhibits spatial structures and northward propagation characteristics broadly similar to observed MISO when confining the analysis to the 25–90 day period band. The MISO precipitation anomalies are then shown to be consistent with previously known observed relationships to broad-scale sea-level pressure patterns, Inter-Tropical Convergence Zone (ITCZ) positioning, and changes in the regional Hadley Cell component. The total simulated seasonal (JJAS) rainfall anomalies over India are not significantly correlated with observations, indicating that intrinsic variations in the regional model atmosphere dominate most of the precipitation variability. However, the bandpass-filtered MISO anomalies surprisingly exhibit a significant correlation (0.61) with observations. This suggests that instabilities in the regional broad-scale atmospheric circulation, e.g., linked to the ITCZ position or strength, may be partly controlled by the large-scale atmospheric flows specified at the domain boundaries and/or that specified local SST anomalies may help to guide some fraction of the developing model MISO to follow observations. This result motivates further research on MISO initiation and development using this type of regional atmospheric model.

The probability of unprecedented high rainfall in wine regions of northern Portugal

Climate is arguably one of the most important factors determining the quality of wine from any given grapevine variety. This study focuses on three wine-growing regions in northern Portugal: Vinho Verde, Trás-os-Montes and Douro, the latter coinciding with Porto. High rainfall during late spring (April to June) can promote growth of the vines but increases the risk of fungal disease. High rainfall during harvest time (August to October) also bears the potential for severe operational disruption and heavy economic losses. The probability of unprecedented rainfall totals in spring and the harvest season over wine-growing regions of northern Portugal has been assessed. A large ensemble of initialised climate model simulations is analysed, and the probability of unprecedented rainfall in each season is quantified. Seasonal rainfall totals considerably higher than any observed are possible in the current climate. An unprecedented rainfall event in either season could occur with a probability between 0.01 and 0.05 in the present climate. Extreme value analysis was applied to rainfall totals from observations and the model ensemble, and the return periods of known extreme rainfall events are calculated. Similar probabilities for unprecedented rainfall totals were calculated. A year similar to 1993, when both seasons were exceptionally wet, would be expected to occur, on average, just once in the next 70–80 years in the current climate. These results could inform the requirements for improved vineyard management and resilience, such as design of drainage channels, access roads and terraces.

Glacier contribution to lowland streamflow: A multi-year, daily geochemical hydrograph separation study in subarctic Alaska

Glacier melt water is a critical fresh water contribution of stream discharge for many regions and in high latitudes glacier extent is decreasing with climate warming. However, characteristics of glacierized contributions to subarctic watershed-scale geochemistry and runoff have attracted limited attention. Tracer studies of glacier dominated watersheds are also commonly limited to a few days. We conducted a six-year geochemical hydrograph separation study of melt season daily streamflow in a glacierized watershed in Interior Alaska to estimate glacier, rain, snow, and baseflow contributions. Bulk streamflow water samples were collected daily from the lowlands of Jarvis Creek (634 km2, 3.3% glacier cover) throughout the flow season (late April through September) during 2011–2016. Synoptic sampling of source waters included rain, snow, headwater late-winter streamflow and late-summer glacier terminus discharge. All 1227 source water and bulk streamflow samples were analyzed for stable water isotopes (δ18O and δD) and dissolved ion concentrations (SO42−:Cl− ratio). Bulk streamflow water samples exhibited large seasonal and inter-annual geochemical variation. Considerable inter-annual differences in stable water isotopes within each source water (δ18O values of −15.8 to −20.9‰ for rain, for example) emphasize the importance of informing the mixing model with source waters sampled during the same season as bulk streamflow collection. We estimated glacier discharge to be the largest seasonal contributor to total warm season lowland streamflow (average 35%, ranging from 20% in year 2016 to 44% in 2012) with daily variability from 2 (June 2, 2011) to 80% (September 10, 2013). If the glacier contribution were to cease completely, total warm season bulk stream discharge would be reduced by between 22 and 48% during an above or below average total seasonal rainfall, respectively, with up to an 80% reduction in peak flows. Results from this study suggest that high variability in source contribution of glacierized watersheds, extending across intra- and interannual timescales, challenge any generalization of hydrologic function derived from short-term studies.