Rainfall Generation Research Articles

A new framework for multi-site stochastic rainfall generator based on empirical orthogonal function analysis and Hilbert-Huang transform

Weather generators (WGs) are tools that create synthetic weather data, which are statistically similar to the observed data. Considering the limitations of most rainfall generators to preserve satisfactorily the fundamental statistical properties (e.g., spatial and temporal coherence) simultaneously, we propose a new framework for multisite rainfall generation. The framework consists of three main components: (1) a spatiotemporal rainfall field, described as spatial modes and their corresponding temporal evolution based on empirical orthogonal function analysis (EOFA). (2) The time series of these spatial modes, decomposed into intrinsic mode functions (IMFs) with characteristic frequencies (periods) using Hilbert-Huang transform (HHT). (3) Stochastic simulation (SS), achieved by assigning random phases for the specific IMFs. The current model, EHS (EOFA + HHT + SS), is compared with two other typical multi-site rainfall generators, MulGETS (parametric model) and KNN (non-parametric model) for a network of 12 stations in Xiang River basin, China. These three models are assessed based on their ability to simulate sequences with statistical attributes that are similar to those observed. We compare the basic statistics (mean, standard deviation, skewness), extreme value characteristics (95th percentile and maximum), spatial dependence (spatial correlation and spatial continuity ratio), and temporal dependence statistic (autocorrelation, wet/dry spells, and low-frequency variability). The results show that EHS rainfall generator has a similar capacity as KNN model in reproducing the spatial structure of the original rainfall field, and has a greater ability than MulGETS and KNN model to preserve the historical temporal statistics, especially the autocorrelation at various time scales and low-frequency variability. Overall, EHS is a useful model for generating realistic multi-site rainfall field and can be expected to generate plausible scenarios for impact studies.

Automatic feature selection and weighting for the formation of homogeneous groups for regional IDF estimation

The intensity-duration frequency (IDF) curve has been used as an effective tool to quantify the risk associated with the impact of extreme rainfall on civil infrastructure. However, recent changes in the rainfall climatology caused by climate change and urbanization have made estimates in the stationary environment provided by the traditional regional IDF approach increasingly inaccurate. This inaccuracy is mainly caused by the lack of consideration for the temporal and spatial differences in the selection of similarity indicators (attributes that are used to measure similarity of extreme rainfall patterns among different stations), resulting in ineffective formation of homogeneous groups (group of stations that share similar extreme rainfall patterns) at various regions. To consider the temporal differences of similarity indicators, including meteorological factors, topographic features and urban impact indicators, a three-layer design is proposed based on the three stages in extreme rainfall formation: cloud formation, rainfall generation and change of rainfall intensity over an urban surface. During the process, the impacts from climate change and urbanization on extreme rainfall patterns are considered through the inclusion of potential features that relate to the rainfall mechanism at each layer. The spatial differences of similarity indicators for Homogeneous Group Formation (HGF) at various regions is resolved by using an automatic feature selection and weighting algorithm, specifically the hybrid searching algorithm of Tabu Search, Lagrange Multiplier and Fuzzy C-means clustering, to select the optimal combination of features for HGF based on the uncertainty in the regional estimates of the rainfall quantiles for a specific site. The proposed methodology fills the gap of including the urbanization impacts on the extreme rainfall patterns during HGF process and challenges the traditional assumption that the same set of features can be equally effective in generating the optimal homogeneous group in regions with different geographic and meteorological characteristics.

Moisture transport and seasonal variations in the stable isotopic composition of rainfall in Central American and Andean Páramo during El Niño conditions (2015–2016)

AbstractHigh‐elevation tropical grassland systems, called Páramo, provide essential ecosystem services such as water storage and supply for surrounding and lowland areas. Páramo systems are threatened by climate and land use changes. Rainfall generation processes and moisture transport pathways influencing precipitation in the Páramo are poorly understood but needed to estimate the impact of these changes, particularly during El Niño conditions, which largely affect hydrometeorological conditions in tropical regions. To fill this knowledge gap, we present a stable isotope analysis of rainfall samples collected on a daily to weekly basis between January 2015 and May 2016 during the strongest El Niño event recorded in history (2014–2016) in two Páramo regions of Central America (Chirripó, Costa Rica) and the northern Andes (Cajas, south Ecuador). Isotopic compositions were used to identify how rainfall generation processes (convective and orographic) change seasonally at each study site. Hybrid Single Particle Lagrangian Integrated Trajectory model (HYSPLIT) air mass back trajectory analysis was used to identify preferential moisture transport pathways to each Páramo site. Our results show the strong influence of north‐east trade winds to transport moisture from the Caribbean Sea to Chirripó and the South American low‐level jet to transport moisture from the Amazon forest to Cajas. These moisture contributions were also related to the formation of convective rainfall associated with the passage of the Intertropical Convergence Zone over Costa Rica and Ecuador during the wetter seasons and to orographic precipitation during the transition and drier seasons. Our findings provide essential baseline information for further research applications of water stable isotopes as tracers of rainfall generation processes and transport in the Páramo and other montane ecosystems in the tropics.

A daily spatially explicit stochastic rainfall generator for a semi-arid climate

Many semi-arid regions of the world experience rainfall patterns characterized by high spatial variability. Accurate spatial representation of different types of rainfall will facilitate the application of distributed hydrological models in these areas. This study presents a daily, spatially distributed, stochastic rainfall generator based on a first-order Markov chain model, calibrated using 50 years of rainfall observations at 88 gages from 1967 through 2016 in the 148-km2 Walnut Gulch Experimental Watershed. Three types of rainfall, including convective, frontal, and tropical depression storms, were simulated separately in the generator using biweekly parameterization. Convective storms were simulated based on an elliptical shape rain cell conceptual model, whereas frontal and tropical depression storms were simulated as uniform rainfall fields over the whole watershed with introduced random variability. The rainfall generator was evaluated by comparing the mean statistics of 30 sets of 50-year simulated data versus the 50-year rain gage observed data. Most individual storm statistics and aggregated seasonal rainfall statistics were similar to the measured rainfall observations. The long-term mean values of both summer and winter rainfall amount were statistically satisfactory. This model can serve as a guide for application in areas with convective, frontal, and tropical depression storms.

Impact of climate variability on hydropower generation in an un-gauged catchment: Erathna run-of-the-river hydropower plant, Sri Lanka

Impact of climate change or climate variability on water resources is an exceedingly concerned issue. Hydropower development is one of the most affected industries due to the climatic variability. Therefore, this paper presents the promising results from a study of the impact of climate variability on hydropower generation of Erathna run-of-the-river (ROR) hydropower plant located in Rathnapura district, Sri Lanka. This study was based on surrounded rain gauges outside the catchment as Erathna catchment area is an un-gauged catchment. 30-year rainfall trend analysis from 1988 to 2017 was done using Mann–Kendall and Sen’s slope estimator tests to predict the available trends. Pearson’s correlation coefficient was used to investigate the relationship between rainfall and Erathna power generation. Results show negative trends for annual rainfalls in several rain gauges, while seasonal trend analyses support that observation. July is the most critical month for most of the rain gauges around the catchment. The results also show a good correlation between the rainfalls and power generation. Therefore, the results conclude the importance of rainfall trend analysis in un-gauged catchments and its forecasting capacity of water resources usage in hydropower development.

A hybrid stochastic rainfall model that reproduces some important rainfall characteristics at hourly to yearly timescales

Abstract. A novel approach to stochastic rainfall generation that can reproduce various statistical characteristics of observed rainfall at hourly to yearly timescales is presented. The model uses a seasonal autoregressive integrated moving average (SARIMA) model to generate monthly rainfall. Then, it downscales the generated monthly rainfall to the hourly aggregation level using the Modified Bartlett–Lewis Rectangular Pulse (MBLRP) model, a type of Poisson cluster rainfall model. Here, the MBLRP model is carefully calibrated such that it can reproduce the sub-daily statistical properties of observed rainfall. This was achieved by first generating a set of fine-scale rainfall statistics reflecting the complex correlation structure between rainfall mean, variance, auto-covariance, and proportion of dry periods, and then coupling it to the generated monthly rainfall, which were used as the basis of the MBLRP parameterization. The approach was tested on 34 gauges located in the Midwest to the east coast of the continental United States with a variety of rainfall characteristics. The results of the test suggest that our hybrid model accurately reproduces the first- to the third-order statistics as well as the intermittency properties from the hourly to the annual timescales, and the statistical behaviour of monthly maxima and extreme values of the observed rainfall were reproduced well.

Climatological moisture sources for the Western North American Monsoon through a Lagrangian approach: their influence on precipitation intensity

Abstract. This work examines the origin of atmospheric water vapor arriving to the western North American monsoon (WNAM) region over a 34-year period (1981–2014) using a Lagrangian approach. This methodology computes budgets of evaporation minus precipitation (E−P) by calculating changes in the specific humidity of thousands of air particles advected into the study area by the observed winds. The length of the period analyzed (34 years) allows the method to identify oceanic and terrestrial sources of moisture to the WNAM region from a climatological perspective. During the wet season, the WNAM region itself is on average the main evaporative source, followed by the Gulf of California. However, water vapor originating from the Caribbean Sea, the Gulf of Mexico, and terrestrial eastern Mexico is found to influence regional-scale rainfall generation. Enhanced (reduced) moisture transport from the Caribbean Sea and the Gulf of Mexico from 4 to 6 days before precipitation events seems to be responsible for increased (decreased) rainfall intensity on regional scales during the monsoon peak. Westward propagating mid- to upper-level inverted troughs (IVs) seem to favor these water vapor fluxes from the east. In particular, a 200 % increase in the moisture flux from the Caribbean Sea to the WNAM region is found to be followed by the occurrence of heavy precipitation in the WNAM area a few days later. Low-level troughs off the coast of northwestern Mexico and upper-level IVs over the Gulf of Mexico are also related to these extreme rainfall events.

Modeling impacts of climate change on return period of landslide triggering

Quantifying the potential influence of climate change on future landslide hazard requires methodologies that allow to properly take into account nonstationarities in the hydro-meteorological causes.In this paper we provide a methodology for estimating return period of landslide triggering under climate change.The methodology capitalizes on the combined use of a stochastic rainfall generator and a hydrological and slope stability model. The stochastic rainfall generator takes into account the statistical dependency between rainfall event duration and intensity through copulas. The hydrological model is based on an analytical solution of a simplified version of the Richards vertical infiltration equation and slope stability is assessed by the infinite slope model. The combined model enables to estimate landslide probability through Monte Carlo simulations. Climate change is then introduced by perturbing the parameters of the rainfall stochastic generator based on factors of change derived from the comparison of future scenarios and the baseline climate as simulated by Regional climate models (RCMs). The Monte Carlo simulations are conducted sequentially on a future moving time window, to derive a yearly series of future landslide triggering probability. This series is then used to compute landslide return period by formulas suitable under nonstationary conditions.An application to the landslide prone region of the Peloritani Mountains, Southern Italy, is carried out to demonstrate the proposed approach. For the application, climate change projections of three RCMs of the MED-CORDEX initiative are used, and a preliminary assessment of the impacts of intermediate- and high-emission Representative concentration pathways (RCPs) 4.5 and 8.5 is carried out.

Similarity in Fog and Rainfall Intermittency

AbstractIntermittent fog occurrences supply significant amounts of moisture to plants in the form of fog drip onto the soil surface thereby prompting interest in their statistical behavior at multiple timescales. A comparison of rainfall and fog measurements collected at an inland tropical cloud forest in Kenya and a coastal rangeland in Northern California is presented to explore whether fog occurrences have similar intermittency characteristics as rainfall. The results confirm that both rainfall and fog show approximate power law relations for distributions of dry period and event size consistent with predictions from self‐organized criticality. Moreover, the spectral exponents of the on‐off time series of the fog and rainfall exhibited an approximate f−0.8 over a broad range of frequencies f, which is remarkably close to scaling exponents across sites experiencing different rainfall generation mechanisms. These results suggest that fog intermittency shares some properties of critical behavior documented in numerous rainfall studies.

Gaps Between the Transmission and Reception of Information on Rainfall Amounts

In weather forecasts, the intensity of rainfall is often expressed either as a quantitative value – the amount of rainfall per hour – or using qualitative language such as “heavy rain.” To date, however, there has been no research into the extent of rainfall that is assumed by information receivers when presented with these qualitative terms. Thus, the present study examines the correspondence between rainfall evaluation and expressions using a rainfall generator. The large-scale rainfall experiment facility owned by the National Research Institute for Earth Science and Disaster Resilience was used to generate rainfall of 60, 180, and 300 mm h-1, and 21 experiment participants experienced this rainfall without knowing the rainfall amounts. Following this, the participants were asked to give feedback using a scale that correlated numerical expressions of rainfall amounts per hour with 10 language expressions such as “heavy rain” and “downpour.” The results revealed that rainfall rates of 60, 180, and 300 mm h-1 were evaluated by the participants as 135, 223, and 311 mm h-1, respectively. The 10 language expressions were felt to be stronger than the official criteria outlined by the Japan Meteorological Agency. In addition, there was no statistical significance among several expressions, suggesting that the qualitative language used to describe different rainfall amounts by information senders were not distinguished by information receivers.

STORM 1.0: a simple, flexible, and parsimonious stochastic rainfall generator for simulating climate and climate change

Abstract. Assessments of water balance changes, watershed response, and landscape evolution to climate change require representation of spatially and temporally varying rainfall fields over a drainage basin, as well as the flexibility to simply modify key driving climate variables (evaporative demand, overall wetness, storminess). An empirical–stochastic approach to the problem of rainstorm simulation enables statistical realism and the creation of multiple ensembles that allow for statistical characterization and/or time series of the driving rainfall over a fine grid for any climate scenario. Here, we provide details on the STOchastic Rainfall Model (STORM), which uses this approach to simulate drainage basin rainfall. STORM simulates individual storms based on Monte Carlo selection from probability density functions (PDFs) of storm area, storm duration, storm intensity at the core, and storm center location. The model accounts for seasonality, orography, and the probability of storm intensity for a given storm duration. STORM also generates time series of potential evapotranspiration (PET), which are required for most physically based applications. We explain how the model works and demonstrate its ability to simulate observed historical rainfall characteristics for a small watershed in southeast Arizona. We explain the data requirements for STORM and its flexibility for simulating rainfall for various classes of climate change. Finally, we discuss several potential applications of STORM.

An Extreme Rainfall Event in Coastal South China During SCMREX‐2014: Formation and Roles of Rainband and Echo Trainings

AbstractIn this study, an extreme rainfall event of 451 mm in 20 hr that occurred in coastal South China on 11 May 2014 during the Southern China Monsoon Rainfall Experiment is investigated using integrated observations from the dual‐Doppler radar pair, polarimetric radar, extensive mesonetwork, and enhanced upper‐air soundings. Results show the generation of the extreme rainfall by two consecutive mesoscale convective systems (MCSs) consisting of multiple meso‐β‐scale rainbands. The two MCSs are maintained by lifting southerly oceanic flows over a quasi‐stationary mesoscale outflow boundary (MOB) along the coastline that are enhanced by convectively generated weak cold pool. Northeastward “echo training” of convective cells, under the influence of environmental southwesterly flows, leads to the formation of the multiple rainbands in each MCS. Their subsequent propagations in a “rainband training” form, together with the echo training, along the coastline account for the production of extreme rainfall. The second MCS is characterized with a leading bowing rainband showing a process of rapid splitting and reestablishment (RSRE), which contributes to the formation of the rainband training. The occurrence of the RSRE process requires ample supply of unstable upstream oceanic air mass, the quasi‐stationary MOB, and a bowing rainband intersecting with the MOB. The second MCS produces more precipitation than the first one as a result of more rainbands, stronger convective intensity, and more moderate‐sized raindrops with larger maximal sizes. The above findings, especially the RSRE process and its associated storm internal circulation, appear to add new Insights into the formation and maintenance of training rainbands and their roles in heavy rainfall production.

Investigating the Plant Species and Rainfall Factors on Stormwater Retention Performance of Extensive Green Roofs in Malaysia

Information on the influences of native plant species and rainfall characteristics on the stormwater retention performance of extensive green roofs in Malaysia is still scarce. More local data is still required because rainfall and runoff generation processes in tropical environment are very different from the temperate regions. This study is aimed to investigate the stormwater retention performance of native plant species in extensive green roofs with respect to different rainfall characteristics in Malaysia. Two test beds consist of pro-mixing potting soil were vegetated each with different native plant species (Axonopus Compressus (cow grass), Portulaca Grandiflora (sedum) and one test bed with no vegetation (bare ground) was prepared as a control. A total of 22 significant storm events were collected over a 3-month period from March 2016 to May 2016. The rainfall depths for monitored storm events were ranged from 4.5 mm to 63.2 mm and rainfall intensities ranged from 0.5mm/hr to 58.5 mm/hr. The results showed that sedum exhibited higher mean runoff retention percentage than grass which is 75.8% and 70.9%, respectively. Bare soil was the least effective for reducing water runoff with retention percentage of 62.6%. Overall, sedum and grass species provided 90.6% and 88.2% of cumulative rainfall retention in this study. The rainfall depth and intensity are correlated negatively with stormwater retention performance of green roofs. Meanwhile, longer dry weather period are likely to increase the water retention capacity of green roof.

Accounting for variation in rainfall intensity and surface slope in wash-off model calibration and prediction within the Bayesian framework

Exponential wash-off models are the most widely used method to predict sediment wash-off from urban surfaces. In spite of many studies, there is still a lack of knowledge on the effect of external drivers such as rainfall intensity and surface slope on wash-off predictions. In this study, a more physically realistic “structure” is added to the original exponential wash-off model (OEM) by replacing the invariant parameters with functions of rainfall intensity and catchment surface slope, so that the model can better represent catchment and rainfall conditions without the need for lookup tables and interpolation/extrapolation. In the proposed new exponential model (NEM), two such functions are introduced. One function describes the maximum fraction of the initial load that can be washed off by a rainfall event for a given slope and the other function describes the wash-off rate during a rainfall event for a given slope. The parameters of these functions are estimated using data collected from a series of laboratory experiments carried out using an artificial rainfall generator, a 1 m2 bituminous road surface and a continuous wash-off measuring system. These experimental data contain high temporal resolution measurements of wash-off fractions for combinations of five rainfall intensities ranging from 33 to 155 mm/h and three catchment slopes ranging from 2 to 8%. Bayesian inference, which allows the incorporation of prior knowledge, is implemented to estimate parameter values. Explicitly accounting for model bias and measurement errors, a likelihood function representative of the wash-off process is formulated, and the uncertainty in the prediction of the NEM is quantified. The results of this study show: 1) even when the OEM is calibrated for every experimental condition, the NEM's performance, with parameter values defined by functions, is comparable to the OEM. 2) Verification indices for estimates of uncertainty associated with the NEM suggest that the error model used in this study is able to capture the uncertainty well.

Influence of uncertain identification of triggering rainfall on the assessment of landslide early warning thresholds

Abstract. Uncertainty in rainfall datasets and landslide inventories is known to have negative impacts on the assessment of landslide-triggering thresholds. In this paper, we perform a quantitative analysis of the impacts of uncertain knowledge of landslide initiation instants on the assessment of rainfall intensity–duration landslide early warning thresholds. The analysis is based on a synthetic database of rainfall and landslide information, generated by coupling a stochastic rainfall generator and a physically based hydrological and slope stability model, and is therefore error-free in terms of knowledge of triggering instants. This dataset is then perturbed according to hypothetical reporting scenarios that allow simulation of possible errors in landslide-triggering instants as retrieved from historical archives. The impact of these errors is analysed jointly using different criteria to single out rainfall events from a continuous series and two typical temporal aggregations of rainfall (hourly and daily). The analysis shows that the impacts of the above uncertainty sources can be significant, especially when errors exceed 1 day or the actual instants follow the erroneous ones. Errors generally lead to underestimated thresholds, i.e. lower than those that would be obtained from an error-free dataset. Potentially, the amount of the underestimation can be enough to induce an excessive number of false positives, hence limiting possible landslide mitigation benefits. Moreover, the uncertain knowledge of triggering rainfall limits the possibility to set up links between thresholds and physio-geographical factors.

Comparison of Gamma and Weibull distributions in simulating hourly rainfall in peninsular Malaysia

Prior knowledge of rainfall is essential in the planning and lessening of the risks associated with extreme events such as floods, landslides and erosions. The development of a reliable model is therefore, crucial in predicting rainfall series that is parallel to the local climate. In this study, a weather generator is chosen to model hourly rainfall time series in Peninsular Malaysia. Short duration rainfall such as hourly scale is an inherent aspect of tropical rainfall, and instigates many flooding events, especially in the western part of the peninsular. Two distributions, Gamma and Weibull are incorporated into the rainfall generator and their performances in generating rainfall series are then compared. Simulations using both Gamma and Weibull distributions are individually conducted at forty stations across the peninsular. Results reveal that both Gamma and Weibull distributions are able to capture rainfall characteristics at the study sites. However, Gamma is found best to represent rainfall at sites located in eastern and southern parts of the peninsular whilst Weibull is more suitable for western and northern parts. Hence, Gamma is more suitable for representing monsoon rainfall while Weibull is more appropriate for inter-monsoon rainfall.

A single scaling parameter as a first approximation to describe the rainfall pattern of a place: application on Catalonia

As well as in other natural processes, it has been frequently observed that the phenomenon arising from the rainfall generation process presents fractal self-similarity of statistical type, and thus, rainfall series generally show scaling properties. Based on this fact, there is a methodology, simple scaling, which is used quite broadly to find or reproduce the intensity–duration–frequency curves of a place. In the present work, the relationship of the simple scaling parameter with the characteristic rainfall pattern of the area of study has been investigated. The calculation of this scaling parameter has been performed from 147 daily rainfall selected series covering the temporal period between 1883 and 2016 over the Catalonian territory (Spain) and its nearby surroundings, and a discussion about the relationship between the scaling parameter spatial distribution and rainfall pattern, as well as about trends of this scaling parameter over the past decades possibly due to climate change, has been presented.

Censored rainfall modelling for estimation of fine-scale extremes

Abstract. Reliable estimation of rainfall extremes is essential for drainage system design, flood mitigation, and risk quantification. However, traditional techniques lack physical realism and extrapolation can be highly uncertain. In this study, we improve the physical basis for short-duration extreme rainfall estimation by simulating the heavy portion of the rainfall record mechanistically using the Bartlett–Lewis rectangular pulse (BLRP) model. Mechanistic rainfall models have had a tendency to underestimate rainfall extremes at fine temporal scales. Despite this, the simple process representation of rectangular pulse models is appealing in the context of extreme rainfall estimation because it emulates the known phenomenology of rainfall generation. A censored approach to Bartlett–Lewis model calibration is proposed and performed for single-site rainfall from two gauges in the UK and Germany. Extreme rainfall estimation is performed for each gauge at the 5, 15, and 60 min resolutions, and considerations for censor selection discussed.

Development and evaluation of a stochastic daily rainfall model with long-term variability

Abstract. The primary objective of this study is to develop a stochastic rainfall generation model that can match not only the short resolution (daily) variability but also the longer resolution (monthly to multiyear) variability of observed rainfall. This study has developed a Markov chain (MC) model, which uses a two-state MC process with two parameters (wet-to-wet and dry-to-dry transition probabilities) to simulate rainfall occurrence and a gamma distribution with two parameters (mean and standard deviation of wet day rainfall) to simulate wet day rainfall depths. Starting with the traditional MC-gamma model with deterministic parameters, this study has developed and assessed four other variants of the MC-gamma model with different parameterisations. The key finding is that if the parameters of the gamma distribution are randomly sampled each year from fitted distributions rather than fixed parameters with time, the variability of rainfall depths at both short and longer temporal resolutions can be preserved, while the variability of wet periods (i.e. number of wet days and mean length of wet spell) can be preserved by decadally varied MC parameters. This is a straightforward enhancement to the traditional simplest MC model and is both objective and parsimonious.

Generation of multi-site stochastic daily rainfall with four weather generators: a case study of Gloucester catchment in Australia

Four weather generators, namely, R-package version of the Generalised Linear Model for daily Climate time series (RGLIMCLIM), Stochastic Climate Library (SCL), R-package multi-site precipitation generator (RGENERATRPREC) and R-package Multi-site Auto-regressive Weather GENerator (RMAWGEN), were used to generate multi-sites stochastic daily rainfall for a small catchment in Australia. The results show the following: (1) All four models produced reasonable results in terms of annual, monthly and daily rainfall occurrence and amount, as well as daily extreme, multi-day extremes and dry/wet spell length. However, they also simulated a large range of variability, which not only demonstrates the advantages of multiple weather generators rather than a single model but also is more suitable for climate change and variability impact studies. (2) Every model has its own advantages and disadvantages due to their different theories and principles. This enhances the benefits of using multiple models. (3) The models can be further calibrated/improved to have a “better” performance in comparison with observations. However, it was chosen not to do so in this case study for two reasons: to obtain a full range of climate variability and to acknowledge the uncertainties associated with observation data. The latter are interpolated from limited stations and therefore have high pairwise correlations—ranging from 0.69 to 0.99 with a median and mean value of 0.87 and 0.88, respectively, for daily rainfall. These conclusions were drawn from a case study in Australia, but could be extended to general guidelines of using weather generators for climate change and variability studies.