S.M. Farmanara; B. Bakhtiari; N. Sayari
Abstract
Introduction: Drought is an extreme climate effect and a creeping phenomenon which directly affects the human life. A drought analysis usually requires characterizing drought severity, duration and frequency (SDF). These characteristic variables are commonly not independent, so this phenomenon is a complex ...
Read More
Introduction: Drought is an extreme climate effect and a creeping phenomenon which directly affects the human life. A drought analysis usually requires characterizing drought severity, duration and frequency (SDF). These characteristic variables are commonly not independent, so this phenomenon is a complex natural disaster and climate change makes it likely to become more frequent and immense in many areas across the world. Therefore, in drought analysis, it is needed to investigate its multivariate nature and spatial variability clearly. Copula, as a model of multivariate distribution, has been used widely in hydrological studies. As the standardized precipitation index (SPI) is more accessible than other indices, it is the most commonly used indicators for analyzing the SDF of meteorological drought. Here, the study has two major focuses: 1) Fitting drought characteristics from SPI to appropriate copulas, then using fitted copulas to estimate conditional drought severity distribution and joint return periods for both historical and future time periods in Fars province. 2) Inquiring the effects of climate change on the frequency and severity of meteorological drought. Materials and Methods: Among the weather stations of Fars province, six synoptic stations were selected, which had longer historical data than others. The data used included 24-hour precipitation during 15 (2004-2018) to 33 (1986-2018) years. Three steps were carried out. Stage one: downscaling of outputs of the large scaling (CanESM2) based on two intermediate (RCP4.5) and pessimistic (RCP8.5) scenarios using SDSM, ver. 4.2.9 during the period of 2020 to 2050. Stage two: calculation of SPIand drought characteristics in the base and future periods (2050-2020). Stage three: extracting SDF curves for the base and future periods under RCP4.5 and RCP8.5 scenarios using copula. The SPIwas used to extract the drought duration and drought severity in the Fars province using GCM models under two selected scenarios (RCP4.5 and RCP8.5) from the IPCC Fifth Assessment Report (AR5) scenarios. The gamble copula function was used to construct the joint distribution function for evaluating the drought return periods in the study area. Because short-term drought prediction is more practical than long-term prediction, we used the 1-month SPI for the copulas-based analysis. Drought severity and duration were calculated based on computed SPIvalues by using the past available data. Drought duration is defined as successive months with SPIvalue less than -1 and drought severity as the accumulative SPIvalue during the period with successive SPIvalue less than -1. The normal and log-normal functions were selected as the candidate distribution function for drought duration and drought severity. Results and Discussion: The results showed that the frequency of drought occurrence in the Fars province will increase during the period of 2020-2050 under the both two scenarios. In the RCP8.5 scenario, the duration of the drought will also increase. The increase and decrease of monthly rainfall in RCP 4.5 and RCP 8.5 were 2.8 and 6.5%, respectively.The duration of the drought were obtained to be 5.25, 5.5 and 6 days at Shiraz station, with a 2 and 5 years return period, in the baseline and future periods under RCP4.5 and RCP8.5 scenarios, respectively. These values were estimated to be 4, 3.5 and 5 days at Bavanat Station.It is expected that the precipitation will decrease at Shiraz station under the two scenarios.Similarly, this amount is expected to increase and decrease at Bavanet station in the RCP4.5 and RCP8.5 scenarios, respectively. Conclusion: Changing droughts based on climate change is important in many aspects. In this study, the performance of two-variable statistical distribution of severity and duration of drought was investigated based on the copula function. The comparison of the drought period calculated using the SPIshowed that due to the climate change, the frequency of drought periods is expected to increase in the base and future periods. The results showed that the value of the precipitation changes in the RCP8.5 scenario is higher than the RCP4.5 scenario. Generally, the performance criteria showed that the SDSM had a good performance for the past and the future periods in Fars province for precipitation data. It is expected that with consideration of the amendments in the sixth report of the IPCC, more precision can be obtained in precipitation modeling. Therefore, reviewing the output of the SDF curves with the availability of the results of this report is suggested.
S.M. Hosseini-Moghari; Sh. Araghinejad
Abstract
Introduction: Due to economic, social, and environmental perplexities associated with drought, it is considered as one of the most complex natural hazards. To investigate the beginning along with analyzing the direct impacts of drought; the significance of drought monitoring must be highlighted. Regarding ...
Read More
Introduction: Due to economic, social, and environmental perplexities associated with drought, it is considered as one of the most complex natural hazards. To investigate the beginning along with analyzing the direct impacts of drought; the significance of drought monitoring must be highlighted. Regarding drought management and its consequences alleviation, drought forecasting must be taken into account (11). The current research employed multi-layer perceptron (MLP), adaptive neuro-fuzzy inference system (ANFIS), radial basis function (RBF) and general regression neural network (GRNN). It is interesting to note that, there has not been any record of applying GRNN in drought forecasting.
Materials and Methods: Throughout this paper, Standard Precipitation Index (SPI) was the basis of drought forecasting. To do so, the precipitation data of Gonbad Kavous station during the period of 1972-73 to 2006-07 were used. To provide short-term, mid-term, and long-term drought analysis; SPI for 1, 3, 6, 9, 12, and 24 months was evaluated. SPI evaluation benefited from four statistical distributions, namely, Gamma, Normal, Log-normal, and Weibull along with Kolmogrov-Smirnov (K-S) test. Later, to compare the capabilities of four utilized neural networks for drought forecasting; MLP, ANFIS, RBF, and GRNN were applied. MLP as a multi-layer network, which has a sigmoid activation function in hidden layer plus linear function in output layer, can be considered as a powerful regressive tool. ANFIS besides adaptive neuro networks, employed fuzzy logic. RBF, the foundation of radial basis networks, is a three-layer network with Gaussian function in its hidden layer, and a linear function in the output layer. GRNN is another type of RBF which is used for radial basis regressive problems. The performance criteria of the research were as follows: Correlation (R2), Root Mean Square Error (RMSE), Mean Absolute Error (MAE).
Results Discussion: According to statistical distribution analysis, the optimal precipitation distribution in many cases was not Gamma distribution. The various time-scales of SPI revealed that, at least in 50% of the events, Gamma was not the selected distribution. Throughout the drought forecasting on the basis of SPI time-series with four aforementioned networks, 80% of the data was allocated to the training process whilst the rest of them considered for the test process. The proper parameters of the networks were chosen via trial and error. Moreover, Cross-validation was used to overcome the over-estimation. The results revealed that the long-term SPIs outdid the others. Performance of the networks promoted with increases in time scales of SPI. In other words, the performance criteria improved proportional to the increases in the time-scales. Based on the Table 3, the least and best performance were contributed to SPI1 and SPI24, respectively. In this regard, R2 of MLP for observed and estimated values of SPI vitiated from 0.009 to 0.949. Similar to MLP, correlation of ANFIS, RBF, and GRNN increased from 0.021 to 0.925, 0.263 to 0.953, and 0.210 to 0.955. Comparison of observed and estimated mean values via Z test indicated that null hypothesis of equal mean observed and estimated values was only rejected for SPI1 with α=0.01. Hence, except SPI1 forecasting, the all other scenarios have remained the mean of observed time series which highlighted the robustness of artificial intelligence in drought forecasting.
Conclusion: The main objective of the ongoing research was monitoring and forecasting of drought based upon various time scales of SPI. In doing so, the precipitation data of Gonbad Kavous station during the period of 1972-73 to 2006-07 were used. Based on K-S test, the best statistical distribution test for different time scales of SPI evaluation was chosen, and then, the SPI was calculated based on the most fitted distribution. After generating the time-series, MLP, ANFIS, RBF, and GRNN were applied for drought forecasting. According to the findings, the lowest performance of forecasting belonged to SPI1 where its RBF’s best performance for R2, RMSE, and MAE were 0.263, 0.806, and 0.989. Furthermore, increases in SPI time-scale promoted the performance of networks. Thus, the worst and best performance belonged to SPI1 and SPI24, respectively. Among the utilized models, ANFIS stood superior to the others, and GRNN followed up after it.
