Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones

STOTEN-135590; No of Pages 13 Science of the Total Environment xxx (xxxx) xxx

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones Myoung-Jin Um a, Yeonjoo Kim b, Daeryong Park c, Kichul Jung c,⁎, Zhan Wang d, Mun Mo Kim e, Hongjoon Shin f a

Department of Civil Engineering, Kyonggi University, Suwon 16227, South Korea Department of Civil and Environmental Engineering, Yonsei University, Seoul 03722, South Korea Department of Civil and Environmental Engineering, Konkuk University, Seoul 05029, South Korea d Finnish Meteorological Institute, 00560 Helsinki, Finland e Department of Civil Engineering, Shingu University, Seongnam 13174, South Korea f Central Research Institute, Korea Hydro & Nuclear Power Co. Ltd., Daejeon 34101, South Korea b c

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• We identify patterns of water deficit and spatial and temporal trends of a drought index. • Two commonly accepted PET estimation methods are analyzed and compared. • Areal extent of spatial trend for drought is increasing in the United States while opposite trends occur in other regions. • Temporal trend of spatial extent for drought is increasing, especially in West Africa for all climate zones.

a r t i c l e

i n f o

Article history: Received 28 May 2019 Received in revised form 29 October 2019 Accepted 15 November 2019 Available online xxxx Editor: Ralf Ludwig Keywords: Drought index Potential evapotranspiration Climate zone Climate Research Unit National Centers for Environmental Prediction

a b s t r a c t The present study is aimed at examining whether potential evapotranspiration (PET), which is important for drought assessment, influences a drought index (standardized precipitation evapotranspiration index; SPEI) for different regions and climate zones. The study regions were East Asia, Europe, the United States (US), and West Africa, and the climate zones considered were the arid, semiarid, subhumid, and humid zones. We examined the pattern of water deficits, spatial trend of the SPEI, area ratio of spatial extent, and temporal trend to provide an understanding of drought characteristics. Two datasets, Climate Research Unit (CRU) and National Centers for Environmental Prediction (NCEP), were used for assessing the drought phenomena. Two types of evapotranspiration obtained using Thornthwaite and Penman–Monteith equations were used to estimate the PET. Negative water deficit values were clearly observed in the arid and semiarid zones of the majority of regions, whereas positive water deficit values were observed in the subhumid and humid zones of the regions. The SPEI spatial trend largely presented a decreasing trend in East Asia and West Africa, a neutral or decreasing trend in Europe, and a neutral or increasing trend in the US. The area ratio of the spatial extent showed large values of a neutral or decreasing trend in East Asia and Europe, a neutral or increasing trend in the US, and a decreasing trend in West Africa. The temporal trend of the spatial extent primarily exhibited no trend or an increasing trend in the aforementioned regions, except in the case of the majority of West Africa. Although the results obtained from the two datasets appear to be slightly different, they show that the PET is predominant in regions,

⁎ Corresponding author. E-mail address: [email protected] (K. Jung).

https://doi.org/10.1016/j.scitotenv.2019.135590 0048-9697/© 2019 Elsevier B.V. All rights reserved.

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

2

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

especially in the US. The PET trends are identified through comparisons and used to understand the drought phenomena while considering various geographic regions and climatic zones. © 2019 Elsevier B.V. All rights reserved.

1. Introduction Generally, drought is defined as a specific period when levels of hydrological variables, including runoff, soil moisture, and precipitation, are lower than the average levels of those variables in a region. This phenomenon negatively affects water resources and the local environment, agriculture, society, and economy (Hagman et al., 1984; Wilhite, 2002; Xu et al., 2015). The determination of the period and severity of drought appears to be difficult because drought is nonstructural and can be spread over large areas of the world. However, the drought phenomenon has been clearly observed, and it has worsened water deficits derived from climate warming (Dai, 2011; Sheffield and Wood, 2008). Sheffield et al. (2012) indicated that the severity of global drought phenomena and long-period drought events has been monitored since the 1970s as global temperatures continued to increase while the precipitation decreased. Furthermore, Ionita et al. (2015) stated that drought events have been widespread in recent years and that these are related to high temperatures and evaporation combined with low precipitation. Several indices based on precipitation, potential evapotranspiration, soil moisture, and other metrics have been used to assess droughts. McKee et al. (1993) and Edwards and McKee (1997) applied the standardized precipitation index (SPI) for identifying classes of droughts on the basis of normalized long-term precipitation. Shukla and Wood (2008) used the standardized runoff index (SRI) based on runoff datasets for determining drought. Moreover, Vicente-Serrano et al. (2010) used the standardized precipitation evapotranspiration index (SPEI) based on the water balance, cumulative water deficit, and loglogistic probability distribution. In a more recent study, Um et al. (2017) also utilized the SPEI with a 12-month lag for the evaluation of drought by investigating historical drought features. Wang et al. (2011) applied the standardized soil water index using soil water datasets for defining the intensity, duration, and frequency of droughts. Zhao and Dai (2017) used the Palmer drought severity index (PDSI) for assessing the drought frequency by investigating the self-calibrated PDSI along with use of the Coupled Model Intercomparison Project Phase 3 and Phase 5 (CMIP3 and CMIP5). In recent years, several studies have been conducted for identifying the effects of potential evapotranspiration with various drought indices on the drought phenomenon. Hao and AghaKouchak (2013) proposed the use of a multivariate drought index based on the concept of copulas to determine the drought features and conditions in California and North Carolina. Zhang et al. (2015) investigated a physically based multi-scalar drought index for providing an improved approach for understanding drought events in the Loess Plateau, which is one of the driest regions in the world. Zhang and He (2016) examined the performance of potential evapotranspiration and the amount of precipitation required to obtain a normal soil-moisture level with the objective of monitoring drought phenomena in arid and semi-arid areas. These researches provided an understanding of the drought features in specific regions with potential evapotranspiration, which we have focused on in this study. In the evaluation of drought, the analysis of the drought trend plays a significant role in the understanding of the drought phenomenon and the management of water resources. Nasrollahi et al. (2015) investigated the accuracy with which global climate models could be used to determine drying and wetting trends based on the Mann–Kendal trend test with the standardized precipitation index (SPI). They used the dataset derived from the Climate Research Unit (CRU) and the prediction of global climate models in CMIP5. Tan et al. (2015) also performed the Mann–Kendal trend test using the SPI and SPEI for

analyzing drought frequency and intensity. They used climate data obtained from 22 meteorological stations in China. Furthermore, Touma et al. (2015) investigated the spatial extent, duration, and drought occurrence based on four drought indices with 15 global climate models in CMIP5. Um et al. (2018) assessed droughts based on the SRI and standardized soil moisture index (SSMI) by evaluating the applicability of the Community Land Model with the CRU dataset. Moreover, the use of various data is important for assessing drought features by estimating a drought index that takes into consideration different potential evaporations. Palmer (1965) used several data including location, precipitation, temperature, and water capacity to obtain the drought index. Dracup et al. (1980) suggested the use of duration, severity, and magnitude of drought for the analysis of historical drought properties. McKee (1995) applied precipitation data to the calculation of water deficit at multiple time scales. Vicente-Serrano et al. (2010) used precipitation and evaporation data to quantify the severity of droughts. Donohue et al. (2010) also applied evaporation and vapor pressure to the assessment of drought by evaluating the availability of potential evaporation formulations. Although the authors used meteorological and geological information for the drought assessment, the effect of various potential evaporations on a drought index has rarely been investigated with the objective of providing a better understanding of the drought phenomenon. The objective of the present study was to investigate the impacts of various types of potential evapotranspiration on the drought index (SPEI) for four representative regions: East Asia, Europe, the United States (US), and West Africa. In addition, climate zones representing arid, semiarid, subhumid, and humid climates were examined for the purpose of comparison. We selected and used the SPEI for the analysis of the drought phenomenon as it has been used widely for hydrological studies as a significant drought index (McEvoy et al., 2012; VicenteSerrano et al., 2011; Zhang et al., 2019; Zhang et al., 2017). The SPEI appears to characterize drought phenomena at regions wherein water consumption increases by taking into consideration the atmospheric evaporative demand (Potop et al., 2012; Scaini et al., 2015). Moreover, the advantage in using the SPEI is that it combines the multi-temporal scalar characteristic of the SPI and its sensitivity to changes in water demand (Vicente-Serrano et al., 2010). Using this drought index, we performed statistical examinations to identify trends in drought phenomena based on various geographic and climatic conditions of the world. Two datasets, the CRU and National Centers for Environmental Prediction (NCEP) datasets, were used to identify the drought phenomenon with the potential evapotranspiration. The analysis was performed by determining the pattern of the monthly average deficits, spatial trend of the SPEI, area ratio of the spatial extent, and temporal trend of the spatial extent of drought. Using the two datasets and various estimations of potential evapotranspiration, the uncertainty in the study of the drought phenomenon can be determined. Using these datasets, two commonly accepted approaches for the estimation of potential evapotranspiration were used and compared in the present work. The two methods include the use of the Thornthwaite equation (Thornthwaite, 1948) and Penman–Monteith equation (Penman, 1948), which are widely applied in the analyses of drought indices (Allen et al., 1998; Sheffield et al., 2012; Zhang et al., 2017). The Thornthwaite equation requires only a few variables such as the daily temperature, day length, and latitude for calculating the potential evapotranspiration. This method, however, tends to overestimate the potential evapotranspiration in humid regions, while it tends to underestimate the potential evapotranspiration in arid and semiarid regions.

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

As the more physically based method, the Penman–Monteith equation can be used to explain the impact of potential evapotranspiration on the drought (McEvoy et al., 2012). Even though the information obtained from the differences can be useful for enhancing the knowledge that can be used to better understand drought characteristics, relatively few studies have dealt with combinations of geographic and climatic conditions while comparing the two representative datasets. The remainder of the present paper is organized as follows. Section 2 presents the materials and methodologies used in this study. Section 3 describes and discusses the results obtained from the analysis of the drought phenomenon. Finally, Section 4 presents the conclusion of this study. 2. Materials and methodologies 2.1. Study region and datasets The study regions of focus are East Asia, Europe, the US, and West Africa in the Northern Hemisphere. These four representative regions have been previously analyzed for identifying drought characteristics (Um et al., 2017). In the present work, two datasets, the CRU and NCEP, were used to investigate the effects of potential evaporation on the drought indices. It should be noted that the NCEP has been used in a reanalysis project with the objective of generating a global dataset with a long-term period and comprising atmospheric parameters. In this study, the period of the temperature and precipitation used to obtain the drought index ranged from 1951 to 2010. Fig. 1 shows the study areas based on the CRU and NCEP for the four considered regions as well as their elevations with latitude and longitude. Furthermore, we classified the climate zones—including arid, semiarid, subhumid, and humid zones—to examine the drought index of the four regions having different climates. Table 1 shows the ranges of precipitation used to define the climate zones. Holdridge (1947) studied plant formations using the various climates to provide an understanding of the impact of the climates on environments. Holdridge (1967) also analyzed the ecology in the different climates based on meteorological and hydrological features. Fig. 2 presents the climate zones in the four regions based on the CRU and NCEP. The CRU dataset that we used for the drought analysis has been developed with aid from several funding sources, including the UK's Natural Environmental Research Council, the US Department of Energy, and the UK National Centre for Atmospheric Science. The dataset, which is a high-resolution gridded set, covers all regions from 60°S to 80°N with a spatial resolution of 0.5°. It includes precipitation, temperature (mean, minimum, maximum, and diurnal), cloud cover, vapor pressure, frost days, rain days, and potential evapotranspiration. Additional details can be found at https://crudata.uea.ac.uk/cru/data/hrg/. The NCEP dataset used for this study is provided by the Earth System Research Laboratory's Physical Sciences Division in the US National

3

Table 1 Ranges of precipitation used to define the arid, semiarid, sub-humid, and humid climate zones. Climate zone

Amount of precipitation (mm/year)

Arid Semiarid Subhumid Humid

b500 500–1000 1000–2000 N2000

Oceanographic and Atmospheric Administration (NOAA). This dataset is an updated globally gridded set with a spatial resolution of 2.5°. It includes several variables, such as temperature, ice concentration, potential evaporation rate, water runoff, precipitation rate, and humidity. The details of the NCEP dataset can be found at https://www.esrl.noaa.gov/ psd/data/gridded/data.ncep.reanalysis.derived.surfaceflux.html. Table 2 shows the two datasets (CRU and NCEP) with the longitude and latitude for the four regions (East Asia, Europe, the US, and West Africa). It should be noted that the longitude and latitude of the regions for the CRU and NCEP are slightly different, because the resolutions of the datasets are different. 2.2. Calculation of drought index There are various indices that can be used for understanding and assessing the drought phenomenon (Jacobi et al., 2013; Nasrollahi et al., 2015; Sheffield et al., 2012; Trenberth et al., 2014). Among these indices, the SPEI has been applied widely for drought analysis using monthly potential evapotranspiration (PET) and monthly precipitation (PR) (Tan et al., 2015; Touma et al., 2015; Um et al., 2017; VicenteSerrano et al., 2010). In the present study, the SPEI was used for the assessment of drought in the four considered regions. The index can be estimated based on the monthly water deficit (D). PET, which is used for the estimation of D, can be obtained using the Thornthwaite equation (TH) (Thornthwaite, 1948) or the Penman–Monteith equation (PM) (Allen, 1994; Harris et al., 2014). It should be noted that the PET characterizes and affects the drought phenomenon as it directly determined the actual atmospheric water demand (Vicente-Serrano et al., 2014; Yang et al., 2016). In the majority of cases, the PET is higher than the evapotranspiration in the nonhumid zone, and the change in the evapotranspiration appears to be dominated by a change in the PET rather than precipitation in the humid zone (Yang et al., 2006; Yang et al., 2007). The PET can be calculated using TH as follows:
PET ¼ 16

L 12





 N 10T d α ; 30 I

ð1Þ

where L is the average length of a day, N is the number of days in a month, and Td is the average daily temperature. α is a constant that

Fig. 1. Study area and elevation. In the figure, the blue box is the range of the CRU dataset and the red box denotes the range of the NCEP dataset. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

4

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

(a) CRU

(b) NCEP Fig. 2. Climate zones, arid, semiarid, subhumid, and humid, analyzed in the present study based on the CRU and NCEP datasets.

D can then be calculated by using the following equation:

can be obtained as follows:       α ¼ 6:75  10−7 I 3 − 7:71  10−5 I2 þ 1:792  10−2 I

D ¼ PR−PET

þ 0:49239;

ð2Þ

where I is a heat index that can be acquired by using Eq. (3).

I¼

 12
X T mi 1:514 i¼1

ð3Þ

5

Once D is calculated, a cumulative difference (Xki, j) over the timescale k in a given year i and month j is estimated to obtain the SPEI. The difference can be obtained as follows: X ki; j ¼

12 X

Di−1;l þ

l¼13−kþ j

X ki; j ¼

Tmi in the above equation is the mean temperature. The PET obtained using PM can be estimated as follows:

ð5Þ

j X

j X

Di;l for jbk

ð6Þ

l¼1

Di;l for j ≥k

ð7Þ

l¼ j−kþ j

To fit the cumulative difference, the following equation is used: 900 U 2 ðea −ed Þ ; PET ¼ 0:408ΔðRn −GÞ þ γ T þ 273:16 Δ þ γ ð1 þ 0:34U 2 Þ

" ð4Þ

where Δ and γ are the slope for the vapor pressure curve and the psychrometric constant, respectively; Rn represents the net radiation; G represents the soil heat flux; (ea − ed) is the vapor pressure deficit; and U2 is the average wind speed at a height of 2 m. Table 2 The four study regions, East Asia, Europe, the United States, and West Africa, based on the CRU and NCEP datasets. Region

East Asia Europe United States West Africa

CRU

EA ER US WA

NCEP

Lon

Lat

Lon

Lat

73.0–135.0 −10.0–40.0 −125.0 to −65.0 −20.0–30.0

20.0–55.0 35.0–70.0 23.0–50.0 0.0–25.0

72.2–135.9 −10.3–40.3 −124.7 to −64.7 −19.7–30.9

19.0–55.2 34.3–70.5 22.8–51.4 0.0–24.8

F ðX Þ ¼ 1 þ




α x−μ

β #−1

;

ð8Þ

where F(X) is the cumulative probability density function with three parameters for D and α, β, and μ are parameters representing the scale, shape, and origin, respectively. The SPEI can then be estimated using the standardized values of the function based on the approximation analyzed by Abramowitz and Stegun (1964). 2.3. Analysis for drought trend For the analysis of regional and climatic trends of the drought phenomenon, the Mann–Kendal trend test (M–K test) was applied to the CRU and NCEP datasets. The M–K test, as a non-parametric test, can be used for statistical assessment when a monotonic upward or monotonic downward trend of variables exists over time (Gilbert, 1987; Kendall, 1948; Mann, 1945). A monotonic upward trend implies that the variables increase consistently over time, while a monotonic downward

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

trend implies that they decrease consistently through time. Increasing or positive trends and decreasing or negative trends were tested at a significance level of 5% in the present study. The spatial trend and trend area ratio of drought were analyzed based on the M–K test. The spatial trend as the spatial extent of the drought can be estimated using the percentage of grid cells that are below a given level of drought in each month. The trend area ratio is the ratio of the grid cells and represents the deceasing, neutral, or increasing trends relative to the total number of grid cells. In this work, we determined the CRU (PR-TH), CRU (PR-PM), NCEP (PR-TH), and NCEP (PR-PM) based on the datasets and methods used for the estimation of the PET. These defined measurements were then applied for the different representative regions (East Asia, Europe, the US, and West Africa) and climatic zones (arid, semiarid, subhumid, and humid) to understand the drought features. We examined the drought trends by identifying a statistically decreasing trend, increasing trend, and no trend at the significance level of 5%. Furthermore, a simple linear regression test was conducted for the temporal changes in the spatial extent of the drought. For the study of the temporal changes, we investigated the temporal changes of the spatial extent of the drought by obtaining the drought trends for the considered regions and climatic zones. 3. Results and discussion 3.1. Pattern of water deficit In the present section, we identify patterns of D based on the CRU and NCEP datasets. D was estimated using PR and PET, which were obtained using the PM and TH. With the datasets and methods used for the estimation of the PET, we defined CRU (PR-TH), CRU (PR-PM), NCEP (PR-TH), and NCEP (PR-PM) for the results. Four regions—East Asia, Europe, the US, and West Africa—were examined in this study. Furthermore, four climate zones—arid, semiarid, subhumid, and humid— were analyzed to provide a better understanding of water deficits. Um et al. (2017) analyzed four regions representing various climatic environments to understand the impact of different reference periods on historical drought phenomena. Van der Schrier et al. (2011) investigated the sensitivity of the PET using the TH and PM methods on the drought characteristics in the arid, semiarid, subhumid, and humid zones. Fig. 3 presents the monthly deficits for the study regions and the four climate zones. In the figure, the combined zone represented as “All” is also included in the analysis in order to determine D for the regions. Fig. 3(a) shows the pattern of D for the four climate zones in East Asia for investigating the effect of the climatic conditions on water deficits. The results of CRU (PR-TH) and CRU (PR-PM) appear to exhibit similar patterns. In the arid and semiarid zones, the values of D are negative, especially from May to September. In contrast, in the subhumid and humid zones, the values of D are positive, especially from May to September. If D has a negative value, the potential evapotranspiration is higher than the precipitation. In this case, the water in the region under study is insufficient, thus resulting in drought. If D has a positive value, the precipitation is higher than the potential evapotranspiration. In this case, the available water in the region under study is sufficient for meeting water demands. The results of NCEP (PR-TH) and NCEP (PRPM) appear to exhibit different patterns. The pattern of NCEP (PR-TH) follows the pattern of CRU (PR-TH) and CRU (PR-PM). However, the pattern of NCEP (PR-PM) tends to be different, as D shows large negative values during nearly all the months. This is because NCEP (PM) shows relatively higher estimations than CRU (TH and PM) and NCEP (TH). As a result, the deficits obtained by NCEP (PR) and NCEP (PM) are lower than those in the other cases, and these conditions then have an effect on the trend of the SPEI in Fig. 4(d). The patterns are also similar to the deficits in other regions such as Europe, the US, and West Africa.

5

Fig. 3(b) presents the pattern of D—similar to that in Fig. 3(a)—for Europe. In the humid zone, CRU (PR-TH) and CRU (PR-PM) present large positive values for all months. In contrast, NCEP (PR-TH) and NCEP (PR-PM) do not have results for D owing to the lack of the dataset for the humid zone and the characteristics of the climate condition in the zone. In the subhumid zone, positive values of D are observed during the majority of the months for CRU (PR-TH), CRU (PR-PM), and NCEP (PR-TH). NCEP (PR-PM) shows a result similar to that of East Asia and presents large negative values for the arid, semiarid, and subhumid climate zones. This may also be due to the fact that the evapotranspiration in NCEP (PR-PM) is quite large, which could affect the result. The results of D with the same datasets for the US are shown in Fig. 3 (c). CRU (PR-TH) and CRU (PR-PM) have positive values of D from October to April and negative values of D from May to September. In particular, large positive values are observed in the humid zone. The results of D appear to be dependent on the climate conditions in the US. However, NCEP (PR-TH) and NCEP (PR-PM) exhibit different patterns. In contrast to CRU (PR-TH) and CRU (PR-PM), NCEP (PR-TH) shows positive values of D from May to September in the humid zone, and NCEP (PR-PM) has large negative values of D in the arid and semiarid zones, which are similar to those of the zones in East Asia and Europe. Fig. 3(d) shows the results of D used for identifying the patterns of the variable in West Africa with the two datasets. The results suggest that the majority of the months have large negative values of D for the arid, semiarid, and subhumid zones for all the datasets, which implies that a severe water shortage exists in West Africa. The results also suggest that the patterns of D depend significantly on the climate conditions in cases wherein the precipitation is low and temperature is high. In particular, NCEP (PR-TH) shows the most negative values in the arid and semiarid zones. In the humid zone, CRU (PR-TH), CRU (PR-PM), and NCEP (PR-TH) have positive values of D, as in the results of the other regions (East Asia, Europe, and the US). 3.2. Spatial trend and trend area ratio We now consider whether the PET has an impact on the spatial trend by using the two datasets (CRU and NCEP) and the two different methods (TH and PM) used to calculate the PET. For the analysis, the M–K trend test was performed for obtaining the SPEI for the four regions. Fig. 4 presents the trend of the SPEI on the basis of the datasets and considered methods in East Asia, Europe, the US, and West Africa. Fig. 4(a) indicates whether the SPEI based on CRU and TH shows a statistically decreasing trend, increasing trend, or no trend at the significance level of 5%. In East Asia, Europe, and West Africa, the areal extent of the decreasing trend is larger than that of the increasing trend. In particular, the decreasing trend is observed in the majority of West Africa. In contrast, the areal extent of the decreasing trend is smaller than that of the increasing trend in the US. Similar to the results of the spatial trend in Fig. 4(a), it can be observed in Fig. 4(b) that the areal extent of the decreasing trend is larger than that of the increasing trend in East Asia and West Africa and that the opposite trend is observed in the US. As shown in this figure, the areal extent of the increasing trend is larger than that of the decreasing trend in Europe. Fig. 4(c) and (d) presents the SPEI trend obtained by using NCEP with TH and that obtained by using NCEP with PM, respectively. In both figures, a large extent of the decreasing trend is observed in East Asia, Europe, and West Africa. In contrast, a large extent of the increasing trend is observed in the US. Moreover, we investigated whether the SPEI has an effect on the area ratio (%) of the trends for the datasets and different methods used in the analysis of the spatial trend. The M–K trend test was also performed in this study, and the results were classified using CRU (PRTH), CRU (PR-PM), NCEP (PR-TH), and NCEP (PR-PM). Fig. 5 shows the area ratio (%) for the spatial trend based on the four zones in the four study regions under consideration. The combined zone is also presented in this figure.

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

6

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

In Fig. 5(a), the area ratio with the analysis of the spatial trend is plotted for East Asia. The values of the area ratio based on neutral trends are highest when using CRU (PR-TH), CRU (PR-PM), NCEP (PR-TH), and NCEP (PR-PM) in the subhumid and humid zones. The neutral trend implies that the climate has a presumably unchanging condition. In the arid and semiarid zones, the majority of the results present decreasing trends, except in the case of CRU (PR-PM) and NCEP (PR-PM) in the arid zone. This might have occurred because the process of estimating the PM affected the area ratio in the zone. The PM equation is a more complex and physically based equation used for estimating the PET. This equation takes into consideration other variables such as vapor pressure deficit, solar radiation, and wind speed that may affect the atmospheric water demand and results of the trends. Furthermore, different trends based on the drought identification obtained from the two

equations for calculating the PET can be observed (Sheffield et al., 2012). Fig. 5(b) shows the area ratio obtained for the same zones in Europe. An increasing trend is observed with CRU (PR-TH) and CRU (PR-PM) in the humid zone, while no result is observed with NCEP (PR-TH) and NCEP (PR-PM) in this zone. This may be a result of a limitation of the NCEP dataset in the humid zone of Europe. CRU (PR-TH) and CRU (PR-PM) show the highest values of neutral trends in the arid, semiarid, and subhumid zones. In contrast, NCEP (PR-TH) and NCEP (PR-PM) show the highest values of decreasing trends in these same zones. In Fig. 5(c), the area ratio for the US is shown for the climate zones. Interestingly, CRU (PR-TH), CRU (PR-PM), NCEP (PR-TH), and NCEP (PR-PM) exhibit only neutral trends in the humid zone. Furthermore, CRU (PR-TH) and CRU (PR-PM) present a highest value of neutral

(a) East Asia

(b) Europe Fig. 3. Monthly average deficits, PR minus PET (TH or PM), depending on the two datasets (CRU and NCEP) and the four climate zones (arid, semiarid, subhumid, and humid).

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

7

(c) United States

(d) West Africa Fig. 3 (continued).

trend in the arid, semiarid, and subhumid zones. In the cases of NCEP, the method of TH provides the highest value of a neutral trend in the arid and semiarid zones, whereas the method of PM provides the highest value of an increasing trend in the arid, semiarid, and subhumid zones. Fig. 5(d) shows the results of the area ratio in West Africa. Most climate zones with the different methods used for estimation of PET show the highest value of a decreasing trend. Only NCEP (PR-TH) and NCEP (PR-PM) in the arid and humid zones exhibit the highest value of a neutral trend.

3.3. Temporal trend of spatial extent Next, we examine whether the PET affects the temporal trend of spatial extent (%) for drought based on the two datasets (CRU and NCEP) and two methods (TH and PM). In all of the analyses described thus far, we provided results based on CRU (PR-TH), CRU (PR-PM), NCEP (PR-TH), and NCEP (PR-PM) for the different climate zones. We also identified the results for the temporal trend of the spatial extent for the same classification in this section. The temporal variations in the

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

8

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

(a) Trend of SPEI with CRU and TH

(b) Trend of SPEI with CRU and PM

(c) Trend of SPEI with NCEP and TH

(d) Trend of SPEI with NCEP and PM Fig. 4. Spatial trend of SPEI depending on the two datasets (CRU and NCEP) and the two types of evapotranspiration (TH and PM).

spatial extent of the SPEI are investigated for the analysis. Fig. 6 presents the results obtained for the climate zones as well as the combined zone. Fig. 6(a) presents the temporal trends for 1951 to 2010 for the purpose of analyzing the spatial extent for the climate zones in East Asia. In this figure, the box plots are shown for each zone. The center line of the box plot indicates the median value for the spatial extent. The bottom and top of the box plot present the 25th and 75th percentiles of the extent, respectively. The dots in the figure indicate outliers. The blue line represents the linear trend derived from a simple linear regression. CRU (PR-TH) and CRU (PR-PM) show an

increasing trend in the spatial extent with the temporal change. NCEP (PR-TH) and NCEP (PR-PM) show a decreasing trend in the humid zone. However, they exhibit an increasing trend in the arid, semiarid, and subhumid zones, except for NCEP (PR-PM) in the arid and subhumid zones. Fig. 6(b) presents the temporal trends of the spatial extent in Europe. CRU (PR-TH) and CRU (PR-PM) show a decreasing trend in the subhumid and humid zones, while they exhibit a slightly increasing trend or no noticeable trend in the arid and semiarid zones. NCEP (PR-TH) and NCEP (PR-PM) present an increasing trend in the arid, semiarid, and subhumid zones. However,

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

9

there are no results obtained for the humid zone of Europe because of the limited availability of the NCEP dataset.

4. Conclusions

In Fig. 6(c), temporal trends are shown for the climate zones in the US. CRU (PR-TH) and CRU (PR-PM) exhibit an increasing trend in the humid zone, where several outliers exist. In other zones, slightly increasing or decreasing trends can be observed. Interestingly, NCEP (PR-TH) and NCEP (PR-PM) present a decreasing trend in all of the climate zones. However, the results of NCEP (PR-TH) and NCEP (PR-PM) in the humid zone seem unreasonable, possibly because limited data availability has affected the results. Fig. 6 (d) shows the temporal trends based on the spatial extent analysis with the climate zones in West Africa. CRU (PR-TH) and CRU (PRPM) clearly exhibit an increasing trend in all of the climate zones. However, NCEP (PR-TH) and NCEP (PR-PM) show an increasing trend in the semiarid and subhumid zones and a decreasing trend in the arid and humid zones. In West Africa, an increasing trend predominates with the drought phenomenon.

This study was aimed at determining how the PET with the SPEI as the drought index affects the drought phenomenon in arid, semiarid, subhumid, and humid climate zones. For the analysis, four different regions (East Asia, Europe, the US, and West Africa) were investigated. The following conclusions were formulated. (i) The patterns of D were determined using CRU (PR-TH), CRU (PRPM), NCEP (PR-TH), and NCEP (PR-PM). In East Asia, the values of D based on CRU (PR-TH) and CRU (PR-PM) were negative for the arid and semiarid zones, while those for the subhumid and humid zones were positive. NCEP (PR-TH) exhibited a pattern similar to those of CRU (PR-TH) and CRU (PR-PM). In Europe, CRU (PR-TH) and CRU (PR-PM) had large positive values. In the US, CRU (PR-TH) and CRU (PR-PM) presented positive patterns for D from October to April. In West Africa, the majority of the results showed large negative values of D for the climate zones.

(a) East Asia

(b) Europe Fig. 5. Trend area ratio (%) of SPEI depending on the two datasets (CRU and NCEP) and the four climate zones (arid, semiarid, subhumid, and humid).

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

10

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

(c) United States

(d) West Africa Fig. 5 (continued).

(ii) The effect of the spatial trend for SPEI was analyzed using the same dataset. The SPEI obtained using CRU and TH in East Asia, Europe, and West Africa indicated that the areal extent of the decreasing trend is larger than that of the increasing trend. In the US, the areal extent of the decreasing trend is smaller than that of the increasing trend. The SPEI based on NCEP with TH and PM showed that a large extent of decreasing trend exists for East Asia, Europe, and West Africa. (iii) The area ratio of the spatial extent was investigated for the drought phenomenon. The majority of the results for East Asia presented decreasing trends in the arid and semiarid zones. In Europe, CRU (PR-TH) and CRU (PR-PM) exhibited the largest values of neutral trends in the arid, semiarid, and subhumid zones. NCEP (PR-TH) and NCEP (PR-PM) exhibited the largest values of decreasing trends in the same zones. In the US, CRU

(PR-TH) and CRU (PR-PM) exhibited the largest values of neutral trends in the arid, semiarid, and subhumid zones. In West Africa, the majority of the results in the climate zones presented the largest value as decreasing trends. (iv) We determined whether the PET has an impact on the temporal trend of spatial extent during the period of 1951 to 2010. In East Asia, CRU (PR-TH) and CRU (PR-PM) exhibited an increasing trend of spatial extent. NCEP (PR-TH) and NCEP (PR-PM) exhibited an increasing trend in the arid, semiarid, and subhumid zones. In Europe, NCEP (PR-TH) and NCEP (PR-PM) showed an increasing trend in the arid, semiarid, and subhumid zones. In the US, CRU (PR-TH) and CRU (PR-PM) presented an increasing trend in the humid zone. In West Africa, an increasing trend was clearly observed for all climate zones with CRU (PR-TH) and CRU (PR-PM).

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

11

(a) East Asia

(b) Europe Fig. 6. Temporal changes of spatial extent of drought, which denotes the area with values less that −1 in SPEI for 1951–2010.

These findings suggest that it is necessary to quantify the drought trends to provide a better understanding of the uncertainty in the atmospheric water demand in various environments. Based on the investigation of PET on a global scale, we can provide well-balanced management of water demand and supply, especially in water-limited regions. Although the use of the TH and PM equations appears to show differences in the determined drought trends, the locations at which PET is predominant can be clearly identified. Furthermore, robust comparisons and evaluations for the drought phenomenon were provided based on

various approaches and datasets in various geographic regions and climatic zones, which can aid in addressing scientific questions regarding droughts. Declaration of competing interest This manuscript has not been published or presented elsewhere in part or in entirety and is not under consideration by another journal. We have read and understood your journal's policies, and we believe

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

12

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx

(c) United States

(d) West Africa Fig. 6 (continued).

that neither the manuscript nor the study violates any of these. There are no conflicts of interest to declare.

Ministry of Science ICT (2019R1F1A1060028) and by Konkuk University Researcher Fund in 2019.

Acknowledgements

References

The authors gratefully acknowledge the helpful comments and suggestions of the two anonymous reviewers to improve the quality of the manuscript. This paper is supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the

Abramowitz, M., Stegun, I.A., 1964. Handbook of Mathematical Functions. Nat. Bur. Stand, Washington, D.C. Allen, R., 1994. An update for the calculation of reference evapotranspiration. ICID Bull. 43 (2), 35–92. Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop evapotranspiration-guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper 300 (9), 56.

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590

M.-J. Um et al. / Science of the Total Environment xxx (xxxx) xxx Dai, A., 2011. Drought under global warming: a review. Wiley Interdiscip. Rev. Clim. Chang. 2 (1), 45–65. Donohue, R.J., McVicar, T.R., Roderick, M.L., 2010. Assessing the ability of potential evaporation formulations to capture the dynamics in evaporative demand within a changing climate. J. Hydrol. 386 (1–4), 186–197. Dracup, J.A., Lee, K.S., Paulson Jr., E.G., 1980. On the definition of droughts. Water Resour. Res. 16 (2), 297–302. Edwards, D.C., McKee, T.B., 1997. Characteristics of 20th century drought in the United States at multiple time scales. Atmospheric Science Paper, USA. Gilbert, R.O., 1987. Statistical Methods for Environmental Pollution Monitoring. John Wiley & Sons. Hagman, G., Beer, H., Bendz, M., Wijkman, A., 1984. Prevention Better than Cure. Report on Human and Environmental Disasters in the Third World. Swedish Red Cross, Stockholm. Hao, Z.C., AghaKouchak, A., 2013. Multivariate standardized drought index: a parametric multi-index model. Adv. Water Resour. 57, 12–18. Harris, I., Jones, P.D., Osborn, T.J., Lister, D.H., 2014. Updated high-resolution grids of monthly climatic observations–the CRU TS3. 10 Dataset. Int. J. Climatol. 34 (3), 623–642. Holdridge, L.R., 1947. Determination of world plant formations from simple climatic data. Science 105 (2727), 367–368. Holdridge, L.R., 1967. Life zone ecology. Tropical Science Center. Ionita, M., Scholz, P., Chelcea, S., 2015. Spatio-temporal variability of dryness/wetness in the Danube River Basin. Hydrol. Process. 29 (20), 4483–4497. Jacobi, J., Perrone, D., Duncan, L.L., Hornberger, G., 2013. A tool for calculating the Palmer drought indices. Water Resour. Res. 49 (9), 6086–6089. Kendall, M.G., 1948. Rank Correlation Methods. Griffin, London. Mann, H.B., 1945. Nonparametric tests against trend. Econometrica 13, 245–259. McEvoy, D.J., Huntington, J.L., Abatzoglou, J.T., Edwards, L.M., 2012. An evaluation of multiscalar drought indices in Nevada and Eastern California. Earth Interact. 16 (18), 1–18. McKee, T.B., 1995. Drought monitoring with multiple time scales. Proceedings of 9th Conference on Applied Climatology, Boston. McKee, T.B., Doesken, N.J., Kleist, J., 1993. The relationship of drought frequency and duration to time scales. Proceedings of the 8th Conference on Applied Climatology. American Meteorological Society Boston. Nasrollahi, N., AghaKouchak, A., Cheng, L., Damberg, L., Phillips, T.J., Miao, C., Hsu, K., Sorooshian, S., 2015. How well do CMIP5 climate simulations replicate historical trends and patterns of meteorological droughts? Water Resour. Res. 51 (4), 2847–2864. Palmer, W., 1965. Meteorological Drought. US Department of Commerce, Weather Bureau, Washington, DC, p. 45 Research Paper. Penman, H.L., 1948. Natural evaporation from open water, bare soil and grass. Proc. R. Soc. Lond. A 193 (1032), 120–145. Potop, V., Možný, M., Soukup, J., 2012. Drought evolution at various time scales in the lowland regions and their impact on vegetable crops in the Czech Republic. Agric. Forest Meteorol. 156, 121–133. Scaini, A., Sánchez, N., Vicente-Serrano, S.M., Martínez-Fernández, J., 2015. SMOS-derived soil moisture anomalies and drought indices: a comparative analysis using in situ measurements. Hydrol. Process. 29 (3), 373–383. Sheffield, J., Wood, E.F., 2008. Projected changes in drought occurrence under future global warming from multi-model, multi-scenario, IPCC AR4 simulations. Clim. Dynam. 31 (1), 79–105. Sheffield, J., Wood, E.F., Roderick, M.L., 2012. Little change in global drought over the past 60 years. Nature 491 (7424), 435–440. Shukla, S., Wood, A.W., 2008. Use of a standardized runoff index for characterizing hydro-

13

logic drought. Geophys. Res. Lett. 35 (2). Tan, C., Yang, J., Li, M., 2015. Temporal-spatial variation of drought indicated by SPI and SPEI in Ningxia Hui Autonomous Region, China. Atmosphere 6 (10), 1399–1421. Thornthwaite, C.W., 1948. An approach toward a rational classification of climate. Geogr. Rev. 38, 55–94. Touma, D., Ashfaq, M., Nayak, M.A., Kao, S.-C., Diffenbaugh, N.S., 2015. A multi-model and multi-index evaluation of drought characteristics in the 21st century. J. Hydrol. 526, 196–207. Trenberth, K.E., Dai, A., Van der Schrier, G., Jones, P.D., Barichivich, J., Briffa, K.P., Sheffield, J., 2014. Global warming and changes in drought. Nat. Clim. Chang. 4 (1), 17–22. Um, M.-J., Kim, Y., Park, D., Kim, J., 2017. Effects of different reference periods on drought index (SPEI) estimations from 1901 to 2014. Hydrol. Earth Syst. Sci. 21 (10). Um, M.-J., Kim, M., Kim, Y., Park, D., 2018. Drought assessment with the community land model for 1951–2010 in East Asia. Sustainability 10 (6). Van der Schrier, G., Jones, P.D., Briffa, K.R., 2011. The sensitivity of the PDSI to the Thornthwaite and Penman-Monteith parameterizations for potential evapotranspiration. J. Geophys. Res. Atmos. 116 (D3). Vicente-Serrano, S.M., Beguería, S., López-Moreno, J.I., 2010. A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index. J. Clim. 23 (7), 1696–1718. Vicente-Serrano, S.M., Lopez-Moreno, J.I., Gimeno, L., Nieto, R., Morán-Tejeda, E., LorenzoLacruz, J., Beguería, S., 2011. A multiscalar global evaluation of the impact of ENSO on droughts. J. Geophys. Res. Atmos. 116 (D20). Vicente-Serrano, S.M., Lopez-Moreno, J.I., Beguería, S., Lorenzo-Lacruz, J., SanchezLorenzo, A., García-Ruiz, J.M., Azorin-Molina, C., Morán-Tejeda, E., Revuelto, J., Trigo, R., Coelho, F., Espejo, F., 2014. Evidence of increasing drought severity caused by temperature rise in southern Europe. Environ. Res. Lett. 9 (4), 044001. Wang, D., Hejazi, M., Cai, X., Valocchi, A.J., 2011. Climate change impact on meteorological, agricultural, and hydrological drought in central Illinois. Water Resour. Res. 47 (9). Wilhite, D.A., 2002. Reducing drought vulnerability through mitigation and preparedness. Report to the Inter-Agency Task Force for Disaster Reduction Sixth Meeting, Geneva. Xu, K., Yang, D., Yang, H., Li, Z., Qin, Y., Shen, Y., 2015. Spatio-temporal variation of drought in China during 1961–2012: a climatic perspective. J. Hydrol. 526, 253–264. Yang, D., Sun, F., Liu, Z., Cong, Z., Lei, Z., 2006. Interpreting the complementary relationship in non-humid environments based on the Budyko and Penman hypotheses. Geophys. Res. Lett. 33 (18). Yang, D., Sun, F., Liu, Z., Cong, Z., Ni, G., Lei, Z., 2007. Analyzing spatial and temporal variability of annual water-energy balance in nonhumid regions of China using the Budyko hypothesis. Water Resour. Res. 43 (4). Yang, Y., Guan, H., Batelaan, O., McVicar, T.R., Long, D., Piao, S., Liang, W., Liu, B., Jin, Z., Simmons, C.T., 2016. Contrasting responses of water use efficiency to drought across global terrestrial ecosystems. Sci. Rep. 6, 23284. Zhang, B., He, C., 2016. A modified water demand estimation method for drought identification over arid and semiarid regions. Agr. For. Meteorol. 230, 58–66. Zhang, B., Zhao, X., Jin, J., Wu, P., 2015. Development and evaluation of a physically based multiscalar drought index: the Standardized Moisture Anomaly Index. J. Geophys. Res. Atmos. 120 (22), 11575–11588. Zhang, B., Wang, Z., Chen, G., 2017. A sensitivity study of applying a two-source potential evapotranspiration model in the Standardized Precipitation Evapotranspiration Index for drought monitoring. Land Degrad. Develop. 28 (2), 783–793. Zhang, B., AghaKouchak, A., Yang, Y., Wei, J., Wang, G., 2019. A water-energy balance approach for multi-category drought assessment across globally diverse hydrological basins. Agr. For. Meteorol. 264, 247–265. Zhao, T., Dai, A., 2017. Uncertainties in historical changes and future projections of drought. Part II: model-simulated historical and future drought changes. Clim. Chang. 144 (3), 535–548.

Please cite this article as: M.-J. Um, Y. Kim, D. Park, et al., Impacts of potential evapotranspiration on drought phenomena in different regions and climate zones, Science of the Total Environment, https://doi.org/10.1016/j.scitotenv.2019.135590