Search

This study evaluated the performance of the climatological precipitation over Southeast Asia using an ensemble of 48 global climate models/earth system models from the sixth phase of Coupled Model Intercomparison Project (CMIP6) suite of experiments. Multi‐source observational data sets were used to evaluate the historical simulation data of the CMIP6 ensembles towards the semi‐year mean of precipitation during November–April (NDJFMA) and May–October (MJJASO). The results indicate that the CMIP6 models were able to reproduce the large‐scale monsoonal features. Precipitation is overestimated over most eastern ocean areas in NDJFMA and over most areas in MJJASO. In terms of the ensemble bias and inter‐model uncertainty, the diagnosis of moisture budget indicates the dominant item is the dynamic convergence and the other important items have dynamic advection and evaporation. The main types of pattern bias were examined through the inter‐model empirical orthogonal function analysis, which is correlated with the global sea surface temperature and moisture flux. The climatological evaluation yielded a higher pattern correlation and a lower root‐mean square error in NDJFMA than MJJASO. Most of the models have greater standard deviations among the spatial grids than observations, especially in NDJFMA. The models with higher resolutions showed a better performance than those with lower resolutions. A model performance index is defined for the convenient comparison, the best three models are CNRM‐CM6‐1, EC‐Earth3‐Veg and NorESM2‐MM, and the worst three are INM‐CM4‐8, AWI‐ESM‐1‐1‐LR and MCM‐UA‐1‐0. This evaluation study provided useful analytics in understanding model behaviour and processes related to rainfall mechanisms which could be further examined in high‐resolution downscaling experiments in future studies. [ABSTRACT FROM AUTHOR]

Large‐scale reforestation can potentially bring both benefits and risks to the water cycle, which needs to be better quantified under future climates to inform reforestation decisions. We identified 477 water‐insecure basins worldwide accounting for 44.6% (380.2 Mha) of the global reforestation potential. As many of these basins are in the Asia‐Pacific, we used regional coupled land‐climate modeling for the period 2041–2070 to reveal that reforestation increases evapotranspiration and precipitation for most water‐insecure regions over the Asia‐Pacific. This resulted in a statistically significant increase in water yield (p <.05) for the Loess Plateau–North China Plain, Yangtze Plain, Southeast China, and Irrawaddy regions. Precipitation feedback was influenced by the degree of initial moisture limitation affecting soil moisture response and thus evapotranspiration, as well as precipitation advection from other reforested regions and moisture transport away from the local region. Reforestation also reduces the probability of extremely dry months in most of the water‐insecure regions. However, some regions experience nonsignificant declines in net water yield due to heightened evapotranspiration outstripping increases in precipitation, or declines in soil moisture and advected precipitation. [ABSTRACT FROM AUTHOR]

A digital elevation model (DEM) represents the topographic surface of the Earth and is an indispensable source of data in many applications, such as flood modeling, infrastructure design and land management. DEM data at high spatial resolution and high accuracy of elevation data are not only costly and time-consuming to acquire but also often confidential. In this paper, we explore a cost-effective approach to derive good quality DEM data by applying a multi-channel convolutional neural network (CNN) to enhance free resources of available DEM data. Shuttle Radar Topography Mission (SRTM) data, multi-spectral imaging Sentinel-2, as well as Google satellite imagery were used as inputs to the CNN model. The CNN model was first trained using high-quality reference DEM data in a dense urban city—Nice, France—then validated on another site in Nice and finally tested in the Orchard Road area (Singapore), which is also an equally dense urban area in Singapore. The CNN model not only shows an impressive reduction in the root mean square error (RMSE) of 50% at validation site in Nice and 30% at the test site in Singapore, but also results in much clearer profiles of the land surface than input SRTM data. A comparison between CNN performance and that of an earlier conducted study using artificial neural networks (ANN) was conducted as well. The comparison within this limited study shows that CNN yields a more accurate DEM. [ABSTRACT FROM AUTHOR]

The study of climate change adaptation plans for drainage infrastructure in a small country as that of Singapore, rainfall projections on the time scale of minutes and on the spatial scales of 1 km are deemed appropriate In this paper, we introduce an application of radar-based stochastic downscaling for rainfall projections at high temporal and spatial resolutions. The input for stochastic model is derived from a Regional Climate Model. The sub-hourly extreme rainfall intensity derived from stochastic model outputs was validated against observed rain-gauge data over the historical period. Considering the advantage in computational efficiency of the stochastic downscaling method, thousand scenarios of rainfall projections at very high temporal and spatial resolution were generated. The implication of this approach is that, from these stochastically downscaled time series of rainfall, it is possible to study future sub-hourly extreme rainfall intensities which would be useful to address issue of flash floods/drainage systems. [ABSTRACT FROM AUTHOR]

The crop simulation model Decision Support System for Agrotechnology Transfer (DSSAT) was applied over the Mekong River Delta (MRD), Southern Vietnam, to assess future (2020–2050) impacts of climate change on rice production. The DSSAT model was driven using observed station data and projected climate data derived through the dynamical downscaling of three global climate models (GCMs) using the Weather Research and Forecasting (WRF) model. The WRF model was simulated at a spatial resolution of 30 km over the study region, and the large-scale driving fields for future climates were taken from the Coupled Model Inter-Comparison Project Phase 3 (CMIP3) global models ECHAM5, CCSM3, and MIROC5 under the A2 emission scenario. Rice growth during two main seasons, namely, the winter-spring (winter) and summer-autumn (summer), were selected to quantify impacts under both irrigated and rain-fed rice cultivation. The results from this climate-crop study suggest that under rain-fed conditions, winter rice yield was likely to experience nearly 24% reduction while summer rice yield was projected to decrease by about 49%. Without irrigation, the annual rice yield was projected to decrease by about 36.5%, and under irrigated conditions, climate change is likely to reduce annual irrigated rice yields by about 1.78%. Winter rice yield was likely to decrease by 4.7% while summer rice yield was projected to marginally increase by about 0.68%. Increasing temperatures and seasonal variations of precipitation are likely to significantly reduce rice yields under rain-fed condition. In addition, (1) a decrease (increase) in the number of rainy days during the dry (wet) season and (2) positive effects of elevated CO2 for rain-fed rice growth under each of the three WRF model realizations would markedly influence rice yields. With Vietnam being one of the largest exporters of rice, these findings have serious implications for the local agricultural sector. This also serves an early warning for the policymakers and stakeholders for effective planning of not only crop production but also water resource management. The findings call for prudent diversification strategy planning by those countries which import rice. [ABSTRACT FROM AUTHOR]

This study evaluates the National Aeronautics Space Administration (NASA) Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) dataset that provides statistically downscaled CMIP5 historical and future climate projections (precipitation and temperature), at high spatial (25 km) and temporal (daily) resolutions. The study is performed over Southeast Asia for the historical period 1976-2005 and compared against gridded observations on daily scales. Future climate change is assessed over the future time slices, 2020-2050 and 2070-2099, under two Representative Concentration Pathways (RCP) scenarios 4.5 and 8.5 with respect to 1976-2005. The future climate projections indicate that surface temperatures over Southeast Asia are likely to increase by more than 3.5 °C by the end of the century. As to precipitation, both the mean and extreme rainfall are likely to increase but the biases in the historical simulations could contribute to larger uncertainties in the estimates of rainfall projections. Findings of the study indicate that NEX-GDDP are in good agreement with observations over the historical period only on monthly scales and that they do not capture the observed statistics on daily scales which suggests that these data need a deeper scrutiny on daily scales, especially when used for impacts studies. [ABSTRACT FROM AUTHOR]

The uncertainties in two high-resolution satellite precipitation products (TRMM 3B42 v7.0 and GSMaP v5.222) were investigated by comparing them against rain gauge observations over Singapore on sub-daily scales. The satellite-borne precipitation products are assessed in terms of seasonal, monthly and daily variations, the diurnal cycle, and extreme precipitation over a 10-year period (2000-2010). Results indicate that the uncertainties in extreme precipitation is higher in GSMaP than in TRMM, possibly due to the issues such as satellite merging algorithm, the finer spatio-temporal scale of high intensity precipitation, and the swath time of satellite. Such discrepancies between satellite-borne and gauge-based precipitations at sub-daily scale can possibly lead to distorting analysis of precipitation characteristics and/or application model results. Overall, both satellite products are unable to capture the observed extremes and provide a good agreement with observations only at coarse time scales. Also, the satellite products agree well on the late afternoon maximum and heavier rainfall of gauge-based data in winter season when the Intertropical Convergence Zone (ITCZ) is located over Singapore. However, they do not reproduce the gauge-observed diurnal cycle in summer. The disagreement in summer could be attributed to the dominant satellite overpass time (about 14:00 SGT) later than the diurnal peak time (about 09:00 SGT) of gauge precipitation. From the analyses of extreme precipitation indices, it is inferred that both satellite datasets tend to overestimate the light rain and frequency but underestimate high intensity precipitation and the length of dry spells. This study on quantification of their uncertainty is useful in many aspects especially that these satellite products stand scrutiny over places where there are no good ground data to be compared against. This has serious implications on climate studies as in model evaluations and in particular, climate model simulated future projections, when information on precipitation extremes need to be reliable as they are highly crucial for adaptation and mitigation. [ABSTRACT FROM AUTHOR]

We present preliminary analyses of the historical (1986-2005) climate simulations of a ten-member subset of the Coupled Model Inter-comparison Project Phase 5 (CMIP5) global climate models over Southeast Asia. The objective of this study was to evaluate the general circulation models’ performance in simulating the mean state of climate over this less-studied climate vulnerable region, with a focus on precipitation. Results indicate that most of the models are unable to reproduce the observed state of climate over Southeast Asia. Though the multi-model ensemble mean is a better representation of the observations, the uncertainties in the individual models are far high. There is no particular model that performed well in simulating the historical climate of Southeast Asia. There seems to be no significant influence of the spatial resolutions of the models on the quality of simulation, despite the view that higher resolution models fare better. The study results emphasize on careful consideration of models for impact studies and the need to improve the next generation of models in their ability to simulate regional climates better. [ABSTRACT FROM AUTHOR]

Optimal designs of stormwater systems rely very much on the rainfall Intensity-Duration-Frequency (IDF) curves. As climate has shown significant changes in rainfall characteristics in many regions, the adequacy of the existing IDF curves is called for particularly when the rainfall are much more intense. For data sparse sites/regions, developing IDF curves for the future climate is even challenging. The current practice for such regions is, for example, to 'borrow' or 'interpolate' data from regions of climatologically similar characteristics. A novel (3-step) Downscaling-Comparison-Derivation (DCD) approach was presented in the earlier study to derive IDF curves for present climate using the extracted Dynamically Downscaled data an ungauged site, Darmaga Station in Java Island, Indonesia and the approach works extremely well. In this study, a well validated (3-step) DCD approach was applied to develop present-day IDF curves at stations with short or no rainfall record. This paper presents a new approach in which data are extracted from a high spatial resolution Regional Climate Model (RCM; 30 × 30 km over the study domain) driven by Reanalysis data. A site in Java, Indonesia, is selected to demonstrate the application of this approach. Extremes from projected rainfall (6-hourly results; ERA40 Reanalysis) are first used to derive IDF curves for three sites (meteorological stations) where IDF curves exist; biases observed resulting from these sites are captured and serve as very useful information in the derivation of present-day IDF curves for sites with short or no rainfall record. The final product of the present-day climate-derived IDF curves fall within a specific range, +38% to +45%. This range allows designers to decide on a value within the lower and upper bounds, normally subjected to engineering, economic, social and environmental concerns. Deriving future IDF curves for Stations with existing IDF curves and ungauged sites with simulation data from RCM driven by global climate model (GCM ECHAM5) (6-hourly results; A2 emission scenario) have also been presented. The proposed approach can be extended to other emission scenarios so that a bandwidth of uncertainties can be assessed to create appropriate and effective adaptation strategies/measures to address climate change and its impacts. Copyright © 2013 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

The overall objective of this study is to improve the forecasting accuracy of the precipitation in the Singapore region by means of both rainfall forecasting and nowcasting. Numerical Weather Predication (NWP) and radar-based rainfall nowcasting are two important sources for quantitative precipitation forecast. In this paper, an attempt to combine rainfall prediction from a high-resolution mesoscale weather model and a radar-based rainfall model was performed. Two rainfall forecasting methods were selected and examined: (i) the weather research and forecasting model (WRF); and (ii) a translation model (TM). The WRF model, at a high spatial resolution, was run over the domain of interest using the Global Forecast System data as initializing fields. Some heavy rainfall events were selected from data record and used to test the forecast capability of WRF and TM. Results obtained from TM and WRF were then combined together to form an ensemble rainfall forecasting model, by assigning weights of 0.7 and 0.3 weights to TM and WRF, respectively. This paper presented results from WRF and TM, and the resulting ensemble rainfall forecasting; comparisons with station data were conducted as well. It was shown that results from WRF are very useful as advisory of anticipated heavy rainfall events, whereas those from TM, which used information of rain cells already appearing on the radar screen, were more accurate for rainfall nowcasting as expected. The ensemble rainfall forecasting compares reasonably well with the station observation data. Copyright © 2012 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

This study used a regional climate model, driven at a resolution of 30 km, to derive climate estimates that were used as input to a hydrological model to determine stream flow in a changing climate. This regional climate model output was derived using the Weather Research and Forecasting model, which was used to downscale the general circulation model ECHAM5 T63 under the A2 greenhouse gas emission scenario for the future. Two river basins, Dakbla and Poko, over the Sesan catchment of the Lower Mekong region were considered for runoff modeling. A 10-year climatology of the recent past, 1991-2000, was used as the baseline for the present-day climate, and another 10-year climate over the period 2091-2100 was chosen for the future time slice. The results from the simulation of future stream flow indicate that, over both Dakbla and Poko river basins, the stream flow is likely to increase, especially during the peak rainfall season. The Dakbla River Basin shows a substantial increase in stream flow when compared with the Poko River Basin. Copyright © 2011 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]