An Application of a Physical Vegetation Model to Estimate Climate Change Impacts on Rice Leaf Wetness

Ryuhei Yoshida Graduate School of Science, Tohoku University, Sendai, Japan

Search for other papers by Ryuhei Yoshida in
Current site
Google Scholar
PubMed
Close
,
Yumi Onodera Graduate School of Science, Tohoku University, Sendai, Japan

Search for other papers by Yumi Onodera in
Current site
Google Scholar
PubMed
Close
,
Takamasa Tojo Graduate School of Science, Tohoku University, Sendai, Japan

Search for other papers by Takamasa Tojo in
Current site
Google Scholar
PubMed
Close
,
Takeshi Yamazaki Graduate School of Science, Tohoku University, Sendai, Japan

Search for other papers by Takeshi Yamazaki in
Current site
Google Scholar
PubMed
Close
,
Hiromitsu Kanno National Agriculture and Food Research Organization, Tsukuba, Japan

Search for other papers by Hiromitsu Kanno in
Current site
Google Scholar
PubMed
Close
,
Izuru Takayabu Meteorological Research Institute, Tsukuba, Japan

Search for other papers by Izuru Takayabu in
Current site
Google Scholar
PubMed
Close
, and
Asuka Suzuki-Parker Graduate School of Life and Environmental Sciences, University of Tsukuba, Japan

Search for other papers by Asuka Suzuki-Parker in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A physical vegetation model [the Two-Layer Model (2LM)] was applied to estimate the climate change impacts on rice leaf wetness (LW) as a potential indicator of rice blast occurrence. Japan was used as an example. Dynamically downscaled data at 20-km-mesh resolution from three global climate models (CCSM4, MIROC5, and MRI-CGCM3) were utilized for present (1981–2000) and future (2081–2100) climates under the representative concentration pathway 4.5 scenario. To evaluate the performance of the 2LM, the LW and other meteorological variables were observed for 108 days during the summer of 2013 at three sites on the Pacific Ocean side of Japan. The derived correct estimation rate was 77.4%, which is similar to that observed in previous studies. Using the downscaled dataset, the changes in several precipitation indices were calculated. The regionally averaged ensemble mean precipitation increased by 6%, although large intermodel differences were found. By defining a wet day as any day in which the daily precipitation was ≥ 1 mm day−1, it was found that the precipitation frequency decreased by 6% and the precipitation intensity increased by 11% for the entire area. The leaf surface environment was estimated to be dry; leaf wetness, wet frequency, and wet times all decreased. It was found that a decrease in water trap opportunities due to reduced precipitation frequency was the primary contributor to the LW decrease. For blast fungus, an increased precipitation intensity was expected to enhance the washout effect on the leaf surface. In the present case, the infection risk was estimated to decrease for Japan.

Current affiliation: Symbiotic Systems Science, Fukushima University, Fukushima, Japan.

Corresponding author address: Ryuhei Yoshida, Faculty of Symbiotic Systems Science, Fukushima University, 1 Kanayagawa, Fukushima 960-1296, Japan. E-mail: yoshida@sss.fukushima-u.ac.jp

Abstract

A physical vegetation model [the Two-Layer Model (2LM)] was applied to estimate the climate change impacts on rice leaf wetness (LW) as a potential indicator of rice blast occurrence. Japan was used as an example. Dynamically downscaled data at 20-km-mesh resolution from three global climate models (CCSM4, MIROC5, and MRI-CGCM3) were utilized for present (1981–2000) and future (2081–2100) climates under the representative concentration pathway 4.5 scenario. To evaluate the performance of the 2LM, the LW and other meteorological variables were observed for 108 days during the summer of 2013 at three sites on the Pacific Ocean side of Japan. The derived correct estimation rate was 77.4%, which is similar to that observed in previous studies. Using the downscaled dataset, the changes in several precipitation indices were calculated. The regionally averaged ensemble mean precipitation increased by 6%, although large intermodel differences were found. By defining a wet day as any day in which the daily precipitation was ≥ 1 mm day−1, it was found that the precipitation frequency decreased by 6% and the precipitation intensity increased by 11% for the entire area. The leaf surface environment was estimated to be dry; leaf wetness, wet frequency, and wet times all decreased. It was found that a decrease in water trap opportunities due to reduced precipitation frequency was the primary contributor to the LW decrease. For blast fungus, an increased precipitation intensity was expected to enhance the washout effect on the leaf surface. In the present case, the infection risk was estimated to decrease for Japan.

Current affiliation: Symbiotic Systems Science, Fukushima University, Fukushima, Japan.

Corresponding author address: Ryuhei Yoshida, Faculty of Symbiotic Systems Science, Fukushima University, 1 Kanayagawa, Fukushima 960-1296, Japan. E-mail: yoshida@sss.fukushima-u.ac.jp

1. Introduction

Rice blast has been recognized as a serious agricultural disease that significantly affects rice yields (Wang et al. 1999). Fungus on a leaf surface can cause a blast outbreak, which has been observed in more than 85 rice-cropped countries (Greer and Webster 2001). For fungus activities, meteorological and soil conditions such as the surface air temperature, water amount on the rice leaf [i.e., leaf wetness (LW)], vapor pressure deficit (VPD), and nitrogen fertilization, are key factors, especially when water is loaded on a rice leaf (Bonman 1992; Luo and Goudriaan 1999; Zhu et al. 2005). In general, an increase of precipitation delivers more water to rice leaves; however, intense precipitation washes out blast fungus, suggesting that the risk of blast infection does not always correspond to the precipitation amount. To understand the risk of blast infection, the precipitation amount and intensity are also essential factors for LW. These precipitation factors combined with LW can provide useful information about the inflectionpotential and may enable farmers to take necessary precautions, such as applying pesticides before blast infections.

Two different methods, empirical and physical models, have been applied in previous LW estimations. As in the cases of other numerical models [e.g., in regional climate models (RCMs); Yoshida et al. 2012], empirical models (e.g., Kim et al. 2002) have been proven to outperform physical models for simulating LW (e.g., Garratt and Segal 1988; Yamazaki et al. 2004), especially when the empirical model parameters have been optimized; however, empirical models require additional calibrations when applied to different spatiotemporal scales (Wichink Kruit et al. 2008; Bregaglio et al. 2011). Because observed LW data are not always available, calibration difficulties exist in empirical models. Moreover, empirical models are based on relationships derived from the present climate; the model should be extrapolated to an assessment of future climate impacts. These difficulties necessitate the application of physical models for future LW simulations at regional or larger scales.

The assessment of climate change impacts on LW has become essential in recent agricultural sectors (Gautam et al. 2013). Because LW causes agricultural diseases, relationships between LW and diseases have been analyzed for various plants, such as tomatoes in America (Kim et al. 2006) and roses in Zimbabwe (Mashonjowa et al. 2013). In the case of rice in Asia, changes in infection risk due to rice blast have been previously estimated by increasing ambient carbon dioxide conditions (Kobayashi et al. 2006) or surface air temperatures (Luo et al. 1998). These studies have suggested that impact assessments of climate change on LW can provide useful information for adaptations. In addition to in situ meteorological variables, the outputs from atmospheric models are now widely used as input data for LW estimation models (Magarey et al. 2006). This trend suggests that we can assess climate change impacts on LW beyond the in situ scale using atmospheric model outputs with a climate change scenario, for example, phase 5 of the Coupled Model Intercomparison Project (CMIP5).

Japan is known for its rice production areas; this region faces significant blast infection risk. In particular, the area in eastern Japan that was damaged by locally intense cold northeasterly winds (known as Yamase) in 1993 and 2003 increased twofold in comparison with that in climatically normal years (i.e., 23.2%; Ministry of Agriculture, Forestry and Fisheries 2014). For a stable food supply, LW relationships with climate change are needed because the LW may be useful for indicating potential for blast infections. To date, one empirical model, the Blast Infection Estimate Model (BLASTAM; Koshimizu 1988; Hayashi and Koshimizu 1988), has been widely used for estimating infection potential. The model first estimates leaf moisture conditions using the Automated Meteorological Data Acquisition System (AMeDAS) and subsequently determines the infection potential through relationships between the estimated leaf moisture condition and the surface air temperature. When evaluating the effects of climate change on LW, BLASTAM encounters many of the aforementioned difficulties that are typical of empirical models. Therefore, physical models are appropriate for such an assessment. One of the physical models, the Two-Layer Model (2LM), calculates the leaf water amount using ambient conditions (e.g., downward shortwave/longwave radiation, air temperature, wind speed, relative humidity, and precipitation). Based on the heat balance equation, the model explicitly calculates dew formation, which is a contributor to LW changes (Luo and Goudriaan 1999). A model that physically calculates LW via the heat balance equation is employed in this study to estimate the climate change effects on LW, in which Japan is used for the analysis region.

Descriptions of the climate change scenarios, physical vegetation model, LW observations, and proposed approach for estimating climate change impacts on LW are included in section 2. Section 3 presents derived performances of the vegetation model and estimated climate change impacts on LW. Relationships between precipitation and LW are discussed in section 4. Advantages and limitations of the proposed approach are also included. Conclusions are presented in section 5.

2. Data and method

a. Climate change scenario

The outputs from CMIP5 global climate models (GCMs) are available to registered members. In this study, we applied the output from the three GCMs: CCSM4 (Gent et al. 2011), MIROC5 (Watanabe et al. 2010), and MRI-CGCM3 (Yukimoto et al. 2012). GCM groups determined using a cluster analysis were used to select the three GCMs (Mizuta et al. 2014; Suzuki-Parker et al. 2014). Mizuta et al. (2014) classified the CMIP5 GCMs into three clusters based on geographical patterns of tropical sea surface temperature (SST) changes. Because of the strong relationship between regional climate in East Asia and tropical SSTs through a teleconnection (e.g., Xie et al. 2010), tropical SSTs can be applied as an index of regional climate over Japan. The three derived GCM clusters were represented by the following characteristics: 1) small SST warming in the South China Sea and Niño region 3, 2) large SST warming in the western North Pacific, and 3) large SST warming in the eastern tropical Pacific. We selected readily available GCMs from each cluster, that is, CCSM4 from cluster 1, MIROC5 from cluster 2, and MRI-CGCM3 from cluster 3.

We applied the GCM outputs for the representative concentration pathway (RCP) 4.5 scenario. Although the CMIP5 output includes the other three RCP scenarios (RCP 2.6, 6.0, and 8.5), we used the scenario that has been most commonly used in GCMs because of limited computational resources (Earth System Grid Federation 2014). Furthermore, 20-yr periods representing both present (1981–2000) and future (2081–2100) climates were used. To clearly detect climate change impacts, we selected the end of the twenty-first century as the future climate period.

The original GCM outputs had a horizontal resolution of 100–150 km, which was too coarse to resolve regional differences in meteorological variables over Japan. Based on this resolution limitation, we conducted a dynamical downscaling of the three GCMs to the 20-km mesh for Japan (Fig. 1) by the Nonhydrostatic Regional Climate Model (NHRCM; Oh’izumi et al. 2013). NHRCM uses the Kain–Fritsch scheme for convective parameterization (Kain 2004) and the improved Mellor–Yamada level 3 turbulence parameterization (Nakanishi and Niino 2004). More details are available in Ishizaki et al. (2012). We simulated a period of 92 days, from 1 June to 31 August (an important period for rice growth), for 20 years for both present (1981–2000) and future (2081–2100) climates.

Fig. 1.
Fig. 1.

Analysis area and calculation domain used in the dynamical downscaling. (a) The domain was 131 × 121 cells with a cell spacing of 20 km for NHRCM. (b) Leaf wetness observation sites: KWT (red dot), FRK (black dot), and KSM (green dot).

Citation: Journal of Applied Meteorology and Climatology 54, 7; 10.1175/JAMC-D-14-0219.1

Using the downscaled precipitation dataset, we calculated the precipitation indices to analyze the relationship between the water supply on the leaf surface and the resulting LW (Table 1). The indices are summarized in Iizumi et al. (2012) for three precipitation indices: the mean precipitation (MEA), the ratio of days with precipitation ≥ 1 mm day−1 (which corresponds to the number of wet days) to the total number of analyzed days (FRE), and the mean precipitation per wet day (INT). The MEA and FRE indices indicate water interception opportunities on the leaf surface, and the remaining index (i.e., INT) provides information regarding decreased blast infection risk due to washout of blast fungus.

Table 1.

Precipitation indices used in this study. Here Na represents the total days during the analysis period (1840 days for the present climate and 1748 days for the future climate), Nw denotes the number of wet days (daily precipitation exceeding 1 mm day−1), Pr indicates the daily precipitation amount, and Pw represents the precipitation amount averaged over the wet days. See Iizumi et al. (2012) for details.

Table 1.

b. Leaf wetness

1) Leaf wetness model

To estimate the intercepted water on the leaf surface, we applied the 2LM physical vegetation model (Yamazaki et al. 1992, 2007; Yamazaki 2001). The 2LM was originally developed to simulate energy–water exchange between plant canopies and the atmosphere. Dividing the rice canopy space into upper and lower layers, the 2LM calculates the leaf water amount in each layer based on six meteorological inputs (i.e., surface air temperature, downward shortwave and longwave radiation, wind speed, relative humidity, and precipitation). Given hourly meteorological inputs, the 2LM calculates the leaf water amount in the ith layer (i = 1, 2), Mi (m), at each time step (Δt = 150 s) as follows:
e1
e2
e3
e4
where Pi is the intercepted precipitation (m s−1), Di denotes the excess water (m s−1), Ei is the evaporation of intercepted water (kg m−2 s−1), ρ is the density of water (kg m−3), PAI signifies the plant area index (m2 m−2), Pr represents the precipitation rate (m s−1), Mx is the maximum water storage on the leaf surface (m), ρa is the air density (kg m−3), ji denotes the leaf evapotranspiration factor (dimensionless), a and amin show the plant area density (m m−2) [=PAI/δ; where δ is thickness of crown space (m)] and its minimum value (=0.0 m m−2), δi is the thickness of each canopy layer (m), ch is the transfer coefficient of individual leaves for sensible heat (dimensionless), Ui is the wind speed (m s−1), qsat represents the saturation specific humidity (kg kg−1), Ti is canopy temperature (K), and qi is the specific humidity (kg kg−1). The maximum water storage on the leaf surface, Mx, is calculated as shown in Kondo et al. (1992):
e5
where sb and sl indicate the storage capacity of the branch and leaf (m), respectively, and PAImin signifies the minimum value of the plant area index (=0.0 m2 m−2). According to the observed rice paddy field PAI data in Yamazaki et al. (1992), we daily increased the PAI value with the same rate from 0.6 on 1 June to 5.0 on 31 August. This PAI setting was applied for all analyzed years.

We first estimate a leaf water amount without consideration of the excess water. If the estimated leaf water amount is greater than the maximum storage, the leaf water amount is set to equal to the maximum value and the difference from the maximum value is defined as the excess water. We analyzed the total intercepted water derived from each rice canopy layer.

2) Observations

From 5 June to 30 September 2013, the selected meteorological variables and leaf surface water were measured at three sites on the Pacific Ocean side of Japan (Fig. 1b). The sites represent mountainous [Kawatabi (KWT), 38°44′36″N, 140°45′36″E; 170 m MSL], inland [Furukawa (FRK), 38°35′54″N, 140°54′42″E; 28 m MSL], and seaside [Kashimadai (KSM), 38°27′36″N, 141°5′30″E; 3 m MSL] areas; all sites were surrounded by rice fields. Using the meteorological and leaf water sensors summarized in Table 2, each element was sampled every 10 min and stored in a datalogger (CR1000, Campbell Scientific).

Table 2.

Summary of the meteorological measurement system.

Table 2.

c. Validation and application of the leaf wetness model

To evaluate the 2LM at simulating leaf water, the observed meteorological variables at the three sites were used as input to the 2LM, and the simulated leaf water amount was analyzed using the observed data. Here, the simulated and observed leaf water information differs because the simulation provides the leaf water amount, whereas the observations provide the leaf water existence. Therefore, the performance of the 2LM was evaluated using the validation table (Table 3) proposed by Wichink Kruit et al. (2008), focusing on water existence on the leaf surface. Four categories were used in the table, namely a “hit,” a “false alarm,” a “miss,” and a “correct rejection.” Two categories were used when both the simulation and the observation exhibited the same situation, that is, a hit (water existence) and a correct rejection (nonexistence). If the simulation (observation) provided the leaf water existence but the observation (simulation) did not, the case was categorized as a false alarm (miss) case.

Table 3.

Validation table for the simulated leaf wetness (see Wichink Kruit et al. 2008 for details).

Table 3.

For assessing climate change impacts on the LW, the downscaled scenario obtained from the NHRCM based on the output from the three GCMs was input into the 2LM. In this evaluation, the analysis area was shifted from the three observation sites to all of Japan (Fig. 1a). When applying meteorological outputs derived from atmospheric models (such as GCMs and RCMs) to impact assessment models, we must be cautious because these outputs are generally biased from the observed values (e.g., Yoshida et al. 2012). To fit the model output to the observed data, various bias correction methods have been developed (e.g., Iizumi et al. 2010). However, considering the six input elements used in the 2LM, the observed downward longwave radiation and/or relative humidity data are unavailable at many sites; thus, the bias correction is difficult for these variables. Limited bias correction for only the variables available in the observed data would lead to physical inconsistencies rather than reducing the bias. Thus, the original downscaled data were used as direct input into the 2LM without any bias correction. Assuming that the model biases are time invariant (so-called constant bias: Buser et al. 2009), the difference between the present and future climates was analyzed to cancel out the biases. However, caution should be taken regarding this assumption because a previous study reported that a linear bias approximation (i.e., constant relation, in which biases depend on the model state) is better than constant bias when considering climate change (Kerkhoff et al. 2014).

3. Result

a. Seasonal changes in the observed meteorological variables

Figure 2 shows the seasonal variations in the meteorological variables observed at the three sites for 108 days (from 5 June to 30 September 2013). Each observed variable was converted from 10-min intervals to daily values.

Fig. 2.
Fig. 2.

Meteorological variables at the three observation sites in 2013: (a) surface air temperature, (b) downward shortwave radiation, (c) relative humidity, and (d) daily precipitation. See Fig. 1 for the locations of each of the three sites.

Citation: Journal of Applied Meteorology and Climatology 54, 7; 10.1175/JAMC-D-14-0219.1

The seasonal surface air temperature variation forms a bell curve, with a maximum value on 10 August for all three sites (KWT: 27.9°C, FRK: 28.2°C, and KSM: 28.4°C; Fig. 2a). Unlike the surface air temperature, the downward shortwave radiation exhibited large daily variations, with continuously decreasing values (Fig. 2b). The maximum during the analyzed period was observed in June, which is common in the study area and has been reported at a site located 60 km south of the three observation sites (Yoshida et al. 2011). The relative humidity exceeded 80% on most of the observed days (KWT: 107 days, FRK: 105 days, and KSM: 100 days; Fig. 2c). Precipitation frequently occurred during the observation period (Fig. 2d), with a maximum on 18 July for the KWT (97.5 mm day−1) and FRK sites (84.0 mm day−1) and on 15 September for the KSM site (75.5 mm day−1).

Table 4 lists the seasonal means of each meteorological variable for the three sites. The surface air temperature the downward shortwave radiation were 1°C and 24% lower, respectively, at the mountainous KWT site than at the FRK and KSM sites. The intersite difference was small for the relative humidity. Precipitation at the KWT site was 36% greater than that at the other two sites.

Table 4.

Warm-season (4 Jun–30 Sep 2013) average meteorological variables at the three observation sites.

Table 4.

b. Verification of the leaf wetness model

Using meteorological variables observed at the three sites in the 2LM, the simulated leaf water was compared with the observations. Among the four categories listed in Table 3, the hit case (category a) had the maximum value for all the sites, followed by the correct rejection case (category d) (Table 5). The sum of the hit and the correct rejection rates corresponds to the performance of the 2LM; the correct ratio was 77.4%. With respect to the failure estimations, the miss case (category c) was distinct for the KWT site, whereas the false alarm case (category b) was evident for both the FRK and the KSM sites. In total, the false alarm rate was larger than that of the miss case.

Table 5.

Statistical scores (%) for the 2LM based on the observation sites. Each term corresponds to that shown in Table 3.

Table 5.

When we focused on wet events (i.e., a case in which leaf water was maintained for at least 1 h), the wet time was generally overestimated, whereas the number of wet events was underestimated expect at the KWT site (Table 6). This finding is consistent with the categorization results (Table 5), that is, the 2LM tends to overestimate water loading. However, the bias fell within a standard deviation of the observed values based on an average of the three sites; thus, the 2LM generally simulated leaf wetness well.

Table 6.

Averaged wet time per wet event for the observation sites. The wet time is expressed as a mean value ± std dev (h); the number of wet events is represented by the total events.

Table 6.

c. Climate change in the lower atmosphere and its impacts on leaf wetness

Figure 3 shows the meteorological variables downscaled to a 20-km mesh using NHRCM based on the three GCM outputs (i.e., CCSM4, MIROC5, and MRI-CGCM3 for the RCP 4.5 scenario). Because no bias correction was applied for each variable and constant bias was assumed, only the differences or ratios between the warm-season average over the 20-yr period for the present (1981–2000) and future (2081–2100) climates are shown in the figure.

Fig. 3.
Fig. 3.

Simulated surface air temperature changes based on (a) CCSM4, (b) MIROC5, and (c) MRI-CGCM3 and (d) the multimodel ensemble for future climate conditions (2081–2100, RCP4.5) relative to the period 1981–2000. (e)–(h) As in (a)–(d), but for daily precipitation changes. Each change is shown as the difference (surface air temperature) or ratio (daily precipitation) of the future climate to the present climate. The bottom-right values indicate the regional averages.

Citation: Journal of Applied Meteorology and Climatology 54, 7; 10.1175/JAMC-D-14-0219.1

The regionally averaged surface air temperature increased by 1.4°–2.5°C, exhibiting a large increase in the northern part of the analysis area based on CCSM4 and MIROC5 and a homogeneous increase based on MRI-CGCM3 (Figs. 3a–c). The ensemble mean of the three GCMs showed a large surface warming in the northern area and less warming in the southern area (Fig. 3d). In contrast to the surface air temperature, the precipitation change exhibited a large intermodel difference in its geographical distribution (Figs. 3e–h). CCSM4 simulated a precipitation decrease over the Sea of Japan side and an increase over the Pacific Ocean side (Fig. 3e), whereas MIROC5 exhibited a reversed distribution, having an opposite sign in 60% of the analyzed grid cells (Fig. 3f). A precipitation increase was widely estimated by MRI-CGCM3, and a large increase was found in the northern area (Fig. 3g). Although the ensemble mean showed an increase in daily precipitation over Japan (6% relative to the present climate), large intermodel differences were found compared to the surface air temperature changes (Fig. 3h).

Next, we calculated the three precipitation indices listed in Table 1. Because the daily precipitation in Figs. 3e–h is identical to the MEA index, the geographical distribution of MEA (Figs. 4a–d) results in the figures that are identical to Figs. 3e–h. The ratio of wet days to the number of analyzed days (i.e., FRE) had a different geographical distribution than MEA, exhibiting a large decrease on the Sea of Japan side based on CCSM4 (Fig. 4e) and on the Pacific Ocean side based on MIROC5 (Fig. 4f). In contrast to the east–west gradient found in CCSM4 and MIROC5, MRI-CGCM3 exhibited a north–south gradient (Fig. 4g). The ensemble mean smoothed intermodel differences in the FRE distribution exhibited a regionally averaged decrease of 6% (Fig. 4h). The intermodel differences were unclear based on the geographical distributions of the precipitation amount on wet days (INT), although all the GCMs simulated an increase in intensity of approximately 10% over the analyzed area (Figs. 4i–l).

Fig. 4.
Fig. 4.

As in Fig. 3, but for changes in the three precipitation indices: (a)–(d) MEA, (e)–(h) FRE, and (i)–(l) INT. The shading and values indicate the ratio of the precipitation index in the future climate (2081–99) to that in the present climate (1981–2000).

Citation: Journal of Applied Meteorology and Climatology 54, 7; 10.1175/JAMC-D-14-0219.1

Figure 5 shows the geographical distributions of leaf wetness estimated using the 2LM with the downscaled climate change scenarios. A decrease in the leaf water amount was estimated for all three GCMs, with area-averaged values ranging from 5% to 19% (Figs. 5a–d). All the GCMs simulated a decrease in leaf wetness in the future climate; however, the decreased area differs in each model. The decrease was distinct on the Sea of Japan side based on CCSM4, the Pacific Ocean side based on MIROC5, and all but the northern part of the analyzed area based on MRI-CGCM3. We then calculated the changes in the number of annual wet events, which was found to decrease in most of the analyzed grid cells (79.2%, 98.0%, and 75.6% of the total number of grid cells in CCSM, MIROC5, and MRI-CGCM3, respectively); the ensemble mean decreased by 7% (Figs. 5e–h). Compared to the LW distribution, regional differences were slightly unclear in the number of wet events. Then, we counted the averaged water loading time (i.e., wet time) per wet event. The wet time was also shortened in many grid cells (39.1, 62.9, and 45.0% in CCSM, MIROC5, and MRI-CGCM3, respectively), with an ensemble mean decrease of 4% (Figs. 5i–l). Although MIROC5 exhibited a large decrease in the ratio of leaf wetness to the number of wet events compared with the other GCMs, the intermodel differences in the regional averages were small for the water loading time, ranging from 2% for CCSM4 to 5% for MIROC5.

Fig. 5.
Fig. 5.

As in Fig. 4, but for changes in leaf wetness variables: (a)–(d) LW, (e)–(h) number of wet events (NWE), and (i)–(l) averaged wet time per wet event (WT). A wet event is defined by the water amount on the leaf surface > 0 mm with a continuous duration ≥ 1 h.

Citation: Journal of Applied Meteorology and Climatology 54, 7; 10.1175/JAMC-D-14-0219.1

Therefore, on the three GCMs (CCSM4, MIROC5, and MRI-CGCM3) with RCP4.5 scenario basis, aridification was estimated for the rice leaf environment (10% decrease in the ensemble mean). However, the mean precipitation, which delivers water to the rice leaf, was predicted to increase over most of the analyzed area. Factors that contributed to the decrease in leaf wetness are discussed in the following session using the precipitation indices.

4. Discussion

a. Impact of climate change on leaf wetness

To identify impactful contributors to the LW decrease, we analyzed changes in the precipitation indices. Figure 6 shows the relationships between the calculated LW changes and the precipitation indices. The mean precipitation and the LW exhibit a positive correlation for all three GCMs, with LW increasing in the presence of more water in most of the grid cells (Figs. 6a–c). However, the slope of the regression line was less than 1; 33% (for CCSM4) to 52% (for MRI-CGCM3) of the increased precipitation contributed to the increase in LW. Because leaf surfaces do not always capture all precipitation due to their capacity depending on the leaf area (i.e., LAI), the existence of untrapped precipitation reduced the slope of the regression line. This untrapped precipitation has been observed in various crops and forests (Dunne and Leopold 1978; Gash et al. 1995; Valente et al. 1997).

Fig. 6.
Fig. 6.

Scatterplot of the normalized leaf wetness and climate changes in the precipitation indices: (a)–(c) MEA, (d)–(f) FRE, and (g)–(i) INT. Each value was normalized as the ratio of the future climate minus the present climate relative to the present climate. Here R2, p, a, and b represent the determination coefficient, the p value, the slope, and the intercept of the regression line, respectively.

Citation: Journal of Applied Meteorology and Climatology 54, 7; 10.1175/JAMC-D-14-0219.1

Furthermore, the precipitation frequency exhibited a positive correlation with the LW (Figs. 6d–f). For all the analyzed GCMs, the correlation coefficient was the largest among the three precipitation indices (with the coefficient being expressed as the determination coefficient of the regression line). Moreover, the slopes were nearly equal to 1 (ranging from 1.00 for CCSM4 to 1.12 for MRI-CGCM3), indicating that the precipitation frequency is directly related to the LW rather than the mean precipitation. Therefore, the simulated climate change impact on the LW was mainly caused by changes in water-supplying opportunities.

In contrast to the previous two indices, there was a weak correlation between the precipitation intensity and the LW (Figs. 6g–i). The determinant coefficient exhibited the lowest value among the precipitation indices, which was a common characteristic for all three GCMs. The existence of untrapped precipitation, as described above, weakened the relationship (Dunne and Leopold 1978; Gash et al. 1995; Valente et al. 1997). However, an increase in the precipitation intensity would decrease the blast infection risk because intense precipitation washes out blast fungus from leaf surfaces.

Because blast fungus activity is controlled by the LW (e.g., Zhu et al. 2005), reduced LW and wet time leads to a decrease in blast infection risk in Japan. Moreover, the increase in the precipitation intensity also reduces risk because of the fungus washout effect. Therefore, based on the leaf water, the future risk would be decreased for Japan when compared with that in the present climate. Globally, the impact of climate change on crop disease infection has previously been shown to exhibit large variations that depend on region, period, and crops (Ghini et al. 2012). For example, the simulated infection risk in France has been predicted to decrease for late blight of potato and grapevine downy mildew in the late twenty-first century because of decreased leaf wet period (Launay et al. 2014), whereas a higher risk has been simulated in the mid-twenty-first century for rice and wheat in Europe (Bregaglio et al. 2013). When surface air temperatures increase, the simulated rice blast fungus becomes active in East Asia (Luo et al. 1998; note that precipitation impacts were excluded in their study). Related to precipitation changes, Chen et al. (2014) suggested that the precipitation increase over the recent five decades has increased the infection risk in humid regions, such as southern China. Although various studies have been conducted for crop diseases, there is a still lack of balance in the literature between diseases (e.g., Launay et al. 2014) and crop yields or growth (e.g., Challinor et al. 2010). This deficiency in the literature is due to difficulties in establishing process-based disease models (Bregaglio et al. 2013). Including a physical model that calculates the LW via the heat balance equation with a disease model would permit a comprehensive study to assess climate change impacts on the agricultural sector.

b. Intermodel differences in the geographical pattern of leaf wetness

The simulated LW changes were found to be sensitive to precipitation, with large intermodel differences between individual GCMs. This discrepancy is a well-known issue in predicting precipitation (Solomon et al. 2007). For Asia, the models in phases 3 and 5 of the Coupled Model Intercomparison Project (known as CMIP3 and CMIP5) exhibited that GCM-dependent precipitation changes because of changes in summer/winter monsoons (Lee and Wang 2014; Ogata et al. 2014, Mizuta et al. 2014). The downscaled climates depend on the parent GCM (e.g., Kawase et al. 2009); therefore, the intermodel variability of GCM climates contributed to the differences in the simulated LW in our study.

Figure 7 shows the geographical patterns of daily precipitation simulated in the three GCMs, corresponding to Figs. 3e–g. General features in the GCM precipitation pattern correspond well with the downscaled precipitation patterns. For example, in MRI-CGCM3, heavy precipitation found in the northern area and the LW increase were simulated in the 2LM using the downscaled climate. Although specific humidity was downscaled in NHRCM instead of precipitation, intermodel differences in the GCMs contributed to the variability in the simulated LW geographical patterns over Japan.

Fig. 7.
Fig. 7.

As in Figs. 3e–g, but for the GCM-simulated changes in daily precipitation based on (a) CCSM4, (b) MIROC5, and (c) MRI-CGCM3. The changes are shown for the future climate (2081–2100) relative to the period 1981–2000. Each change is shown as the ratio of the future climate to the present climate.

Citation: Journal of Applied Meteorology and Climatology 54, 7; 10.1175/JAMC-D-14-0219.1

c. Advantages and remaining uncertainties in the leaf wetness model

The correct estimation ratio of the 2LM was 77.4% for the 324 site-days (108 days × 3 sites). This performance is comparable to that of other physical models evaluated by Wichink Kruit et al. (2008): 78% for Garratt and Segal (1988), 75% for Pedro and Gillespie (1981a,b), and 85% for Monteith (1981). The former two models estimated the potential dew from VPD and the net radiation, whereas the latter model only used the net radiation. The models were applied for grass with a height of 0.1 m. Although their validation period was limited to 39 days and the analyzed vegetation was different from that in our study, a multilayer treatment would be needed for relatively tall vegetation (maximum height of 0.9 m) to maintain the performance of the physical model in simulating LW as a level similar to that of grass.

Among the validation statistics presented in Table 5, the 2LM tends to overestimate the leaf water existence. To improve the model performance, observation data are needed, especially for the storage capacity of the leaf surface. Moreover, the observed data were limited to the three sites and one summer in this study; thus, long-term and widespread observations would improve the 2LM through validations from various aspects.

5. Conclusions

This study applied a physical vegetation model (the 2LM) to estimate climate change impacts on the LW for Japan using a climate change dataset derived from three GCMs (CCSM4, MIROC5, and MRI-CGCM3) with the RCP 4.5 scenario. There was a strong correlation between the LW and precipitation; the LW decreased because of reduced precipitation frequency. Moreover, the simulated future precipitation was intensified throughout the area. Because blast fungus would be more susceptible to washout because of the intensified precipitation, the LW decrease and intensified precipitation would suppress the blast infection risk in most of Japan based on the three GCMs and RCP 4.5 scenario presented herein.

The correct estimation ratio of the 2LM was 77.4% for the site observations, which was found to be comparable to those of other LW models validated for grass surfaces. In addition, the 2LM tended to overestimate the leaf water loading. Observed data, such as the storage capacity on the leaf surface, would be useful in improving the model.

The estimated LW change was used to determine the blast infection potential. Therefore, impact assessments focusing on rice blast itself are needed to connect the relationship between the LW and rice blast infections. Furthermore, to achieve a stable rice yield in the future, rice blast is an aspect that should be considered; thus, comprehensive impact assessments, including high/low temperature injuries or pest damages are essential for food security.

Acknowledgments

This study was supported by JSPS KAKENHI (Grants 24540467, 25892004, and 26350412), Asahi Group Foundation, the Research Program on Climate Change Adaptation (RECCA) and the Program for Risk Information on Climate Change (SOUSEI) funded by the Ministry of Education, Culture, Sports, Science, and Technology of Japan, and the Cross-ministerial Strategic Innovation Promotion Program of the Cabinet Office, Government of Japan. We thank Mr. Hiroei Kanno, Dr. Sayuri Okubo, Dr. Shota Ishii, Mr. Sho Ikeda, and the staffs of the Kashimadai Experimental Farm Station of the Institute of Genetic Ecology and Tohoku University Kawatabi Field Center for their operational support; Dr. Toshinori Aoyagi supported the data operation for the climate change scenario. Some of the experimentally obtained data in this research were obtained using the supercomputing resources of the Cyberscience Center at Tohoku University. Our thanks are extended to the editor and three anonymous reviewers for their many informative and invaluable comments.

REFERENCES

  • Bonman, J. M., 1992: Durable resistance to rice blast disease—Environmental influences. Euphytica, 63, 115123, doi:10.1007/BF00023917.

    • Search Google Scholar
    • Export Citation
  • Bregaglio, S., M. Donatelli, R. Confalonieri, M. Acutis, and S. Orlandini, 2011: Multi metric evaluation of leaf wetness models for large-area application of plant disease models. Agric. For. Meteor., 151, 11631172, doi:10.1016/j.agrformet.2011.04.003.

    • Search Google Scholar
    • Export Citation
  • Bregaglio, S., M. Donatelli, and R. Confalonieri, 2013: Fungal infections of rice, wheat, and grape in Europe in 2030–2050. Agron. Sustainable Dev., 33, 767776, doi:10.1007/s13593-013-0149-6.

    • Search Google Scholar
    • Export Citation
  • Buser, C. M., H. R. Künsch, D. Lüthi, M. Wild, and C. Schär, 2009: Bayesian multi-model projection of climate: Bias assumption and interannual variability. Climate Dyn., 33, 849868, doi:10.1007/s00382-009-0588-6.

    • Search Google Scholar
    • Export Citation
  • Challinor, A. J., E. S. Simelton, E. D. G. Fraser, D. Hemming, and M. Collins, 2010: Increased crop failure due to climate change: Assessing adaptation options using models and socio-economic data for wheat in China. Environ. Res. Lett., 5, 034012, doi:10.1088/1748-9326/5/3/034012.

    • Search Google Scholar
    • Export Citation
  • Chen, C., G.-S. Zhou, and L. Zhou, 2014: Impacts of climate change on rice yield in China from 1961 to 2010 based on provincial data. J. Integr. Agric., 13, 15551564, doi:10.1016/S2095-3119(14)60816-9.

    • Search Google Scholar
    • Export Citation
  • Dunne, T., and L. N. Leopold, 1978: Water in Environmental Planning. W. H. Freeman and Company, 818 pp.

  • Earth System Grid Federation, 2014: ESGF portal. Program for Climate Model Diagnosis and Intercomparison, accessed 25 Aug 2014. [Available online at http://pcmdi9.llnl.gov/esgf-web-fe/.]

  • Garratt, J. R., and M. Segal, 1988: On the contribution of atmospheric moisture to dew formation. Bound.-Layer Meteor., 45, 209236, doi:10.1007/BF01066671.

    • Search Google Scholar
    • Export Citation
  • Gash, J. H. C., C. R. Lloyd, and G. Lachaud, 1995: Estimating sparse forest rainfall interception with an analytical model. J. Hydrol., 170, 7986, doi:10.1016/0022-1694(95)02697-N.

    • Search Google Scholar
    • Export Citation
  • Gautam, H. R., M. L. Bhardwaj, and R. Kumar, 2013: Climate change and its impact on plant diseases. Curr. Sci., 105, 16851691.

  • Gent, P. R., and Coauthors, 2011: The Community Climate System Model version 4. J. Climate, 24, 49734991, doi:10.1175/2011JCLI4083.1.

    • Search Google Scholar
    • Export Citation
  • Ghini, R., E. Hamada, F. Angelotti, L. B. Costa, and W. Bettiol, 2012: Research approaches, adaptation strategies, and knowledge gaps concerning the impacts of climate change on plant diseases. Trop. Plant Pathol., 37 (1), 524.

    • Search Google Scholar
    • Export Citation
  • Greer, C. A., and R. K. Webster, 2001: Occurrence, distribution, epidemiology, cultivar reaction, and management of rice blast disease in California. Plant Dis., 85, 10961102, doi:10.1094/PDIS.2001.85.10.1096.

    • Search Google Scholar
    • Export Citation
  • Hayashi, T., and Y. Koshimizu, 1988: Computer program BLASTAM for forecasting occurrence of rice leaf blast (in Japanese). Bull. Tohoku Natl. Agric. Exp. Stn., 78, 123138.

    • Search Google Scholar
    • Export Citation
  • Iizumi, T., M. Nishimori, Y. Ishigooka, and M. Yokozawa, 2010: Introduction to climate change scenario derived by statistical downscaling (in Japanese with English abstract). J. Agric. Meteor., 66 (2), 131143, doi:10.2480/agrmet.66.2.5.

    • Search Google Scholar
    • Export Citation
  • Iizumi, T., and Coauthors, 2012: Future change of daily precipitation indices in Japan: A stochastic weather generator-based bootstrap approach to provide probabilistic climate information. J. Geophys. Res., 117, D11114, doi:10.1029/2011JD017197.

    • Search Google Scholar
    • Export Citation
  • Ishizaki, N. N., and Coauthors, 2012: Improved performance of simulated Japanese climate with a multi-model ensemble. J. Meteor. Soc. Japan, 90, 235254, doi:10.2151/jmsj.2012-206.

    • Search Google Scholar
    • Export Citation
  • Kain, J. S., 2004: The Kain–Fritch convective parameterization: An update. J. Appl. Meteor., 43, 170181, doi:10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kawase, H., T. Yoshikane, M. Hara, B. Ailikun, F. Kimura, T. Yasunari, T. Inoue, and H. Ueda, 2009: Intermodel variability of future changes in the Baiu rainband estimated by the pseudo global warming downscaling method. J. Geophys. Res., 114, D24110, doi:10.1029/2009JD011803.

    • Search Google Scholar
    • Export Citation
  • Kerkhoff, C., H. Künsch, and C. Schär, 2014: Assessment of bias assumptions for climate models. J. Climate, 27, 67996818, doi:10.1175/JCLI-D-13-00716.1.

    • Search Google Scholar
    • Export Citation
  • Kim, K. S., S. E. Taylor, M. L. Gleason, and K. J. Koehler, 2002: Model to enhance site-specific estimation of leaf wetness duration. Plant Dis., 86, 179185, doi:10.1094/PDIS.2002.86.2.179.

    • Search Google Scholar
    • Export Citation
  • Kim, K. S., M. L. Gleason, and S. E. Taylor, 2006: Forecasting site-specific leaf wetness duration for input to disease-warning systems. Plant Dis., 90, 650656, doi:10.1094/PD-90-0650.

    • Search Google Scholar
    • Export Citation
  • Kobayashi, T., K. Ishiguro, T. Nakajima, H. Y. Kim, M. Okada, and K. Kobayashi, 2006: Effects of elevated atmospheric CO2 concentration on the infection of rice blast and sheath blight. Ecol. Epidemiol., 96, 425431.

    • Search Google Scholar
    • Export Citation
  • Kondo, J., T. Watanabe, M. Nakazono, and M. Ishii, 1992: Estimation of forest rainfall interception (in Japanese). Tenki, 39, 159167.

    • Search Google Scholar
    • Export Citation
  • Koshimizu, Y., 1988: A forecasting method for occurrence of rice leaf blast with AMeDAS data (in Japanese). Bull. Tohoku Natl. Agric. Exp. Stn., 78, 67121.

    • Search Google Scholar
    • Export Citation
  • Launay, M., J. Caubel, G. Bourgeois, F. Huard, I. G. Cortazar-Atauri, M.-O. Bancal, and N. Brisson, 2014: Climatic indicators for crop infection risk: Application to climate change impacts on five major foliar fungal diseases in northern France. Agric. Ecosyst. Environ., 197, 147158, doi:10.1016/j.agee.2014.07.020.

    • Search Google Scholar
    • Export Citation
  • Lee, J.-Y., and B. Wang, 2014: Future change of global monsoon in the CMIP5. Climate Dyn., 42, 101119, doi:10.1007/s00382-012-1564-0.

    • Search Google Scholar
    • Export Citation
  • Luo, W., and J. Goudriaan, 1999: Effects of altering water temperature on leaf wetness in paddy rice crops. Agric. For. Meteor., 97, 3342, doi:10.1016/S0168-1923(99)00046-5.

    • Search Google Scholar
    • Export Citation
  • Luo, Y., P. S. Teng, N. G. Febellar, and D. O. TeBeest, 1998: The effects of global temperature change on rice leaf blast epidemics: A simulation study in three agroecological zones. Agric. Ecosyst. Environ., 68, 187196, doi:10.1016/S0167-8809(97)00082-0.

    • Search Google Scholar
    • Export Citation
  • Magarey, R. D., J. M. Russo, and R. C. Seem, 2006: Simulation of surface wetness with a water budget and energy balance approach. Agric. For. Meteor., 139, 373381, doi:10.1016/j.agrformet.2006.08.016.

    • Search Google Scholar
    • Export Citation
  • Mashonjowa, E., F. Ronsse, M. Mubvuma, J. R. Milford, and J. G. Pieters, 2013: Estimation of leaf wetness duration for greenhouse roses using a dynamic greenhouse climate model in Zimbabwe. Comput. Electron. Agric., 95, 7081, doi:10.1016/j.compag.2013.04.007.

    • Search Google Scholar
    • Export Citation
  • Ministry of Agriculture, Forestry and Fisheries, 2014: Crop statistics (in Japanese). Accessed 12 May 2014. [Available online at http://www.maff.go.jp/j/tokei/kouhyou/sakumotu/sakkyou_kome/index.html#r.]

  • Mizuta, R., O. Arakawa, T. Ose, S. Kusunoki, H. Endo, and A. Kitoh, 2014: Classification of CMIP5 future climate responses by the tropical sea surface temperature changes. SOLA,10, 167–171, doi:10.2151/sola.2014-035.

  • Monteith, J. L., 1981: Evaporation and surface temperature. Quart. J. Roy. Meteor. Soc., 107, 127, doi:10.1002/qj.49710745102.

  • Nakanishi, M., and H. Niino, 2004: An improved Mellor–Yamada level-3 model with condensation physics: Its design and verification. Bound.-Layer Meteor., 112, 131, doi:10.1023/B:BOUN.0000020164.04146.98.

    • Search Google Scholar
    • Export Citation
  • Ogata, T., H. Ueda, T. Inoue, M. Hayashi, A. Yoshida, S. Watanabe, M. Kira, M. Ooshiro, and A. Kumai, 2014: Projected future changes in the Asian Monsoon: A comparison of CMIP3 and CMIP5 model results. J. Meteor. Soc. Japan, 92, 207225, doi:10.2151/jmsj.2014-302.

    • Search Google Scholar
    • Export Citation
  • Oh’izumi, M., I. Takayabu, and N. Ishizaki, 2013: Reproducibility of snow by a non-hydrostatic regional climate model (NHRCM) nested in JRA-25. J. Meteor. Soc. Japan, 91, 229238, doi:10.2151/jmsj.2013-209.

    • Search Google Scholar
    • Export Citation
  • Pedro, M. J., and T. J. Gillespie, 1981a: Estimating dew formation. I. Utilizing micrometeorological data. Agric. Meteor., 25, 283296, doi:10.1016/0002-1571(81)90081-9.

    • Search Google Scholar
    • Export Citation
  • Pedro, M. J., and T. J. Gillespie, 1981b: Estimating dew formation. II. Utilizing standard weather station data. Agric. Meteor., 25, 297310, doi:10.1016/0002-1571(81)90082-0.

    • Search Google Scholar
    • Export Citation
  • Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K. Averyt, M. Tignor, and H. L. Miller Jr., Eds., 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

  • Suzuki-Parker, A., I. Takayabu, R. Mizuta, H. Kusaka, K. Dairaku, S. Ham, S. A. Adachi, and N. N. Ishizaki, 2014: An overview of the SOUSEI multi GCM/RCM dynamical downscaling ensemble. Proc. 21st Century Challenges in Regional Climate Modelling, Lund, Sweden, International BALTEX Secretariat, 374–375. [Available online at http://www.baltex-research.eu/RCM2014/Material/IBESPS_No.3_low.pdf.]

  • Valente, F., J. S. David, and J. H. C. Gash, 1997: Modelling interception loss for two sparse eucalypt and pine forests in central Portugal using reformulated Rutter and Gash analytical models. J. Hydrol., 190, 141162, doi:10.1016/S0022-1694(96)03066-1.

    • Search Google Scholar
    • Export Citation
  • Wang, Z.-X., M. Yano, U. Yamanouchi, M. Iwamoto, L. Monna, H. Hayasaka, Y. Katayose, and T. Sasaki, 1999: The Pib gene for rice blast resistance belongs to the nucleotide binding and leucine-rich repeat class of plant disease resistance genes. Plant J., 19, 5564, doi:10.1046/j.1365-313X.1999.00498.x.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., and Coauthors, 2010: Improved climate simulation by MIROC5: Mean states, variability, and climate sensitivity. J. Climate, 23, 63126335, doi:10.1175/2010JCLI3679.1.

    • Search Google Scholar
    • Export Citation
  • Wichink Kruit, R. J., A. F. G. Jacobs, and A. A. M. Holtslag, 2008: Measurements and estimates of leaf wetness over agricultural grassland for dry deposition modeling of trace gases. Atmos. Environ., 42, 53045316, doi:10.1016/j.atmosenv.2008.02.061.

    • Search Google Scholar
    • Export Citation
  • Xie, S.-P., C. Deser, G. A. Vecchi, J. Ma, H. Teng, and A. T. Wittenberg, 2010: Global warming pattern formation: Sea surface temperature and rainfall. J. Climate, 23, 966986, doi:10.1175/2009JCLI3329.1.

    • Search Google Scholar
    • Export Citation
  • Yamazaki, T., 2001: A one-dimensional land surface model adaptable to intensely cold regions and its applications in eastern Siberia. J. Meteor. Soc. Japan, 79, 11071118, doi:10.2151/jmsj.79.1107.

    • Search Google Scholar
    • Export Citation
  • Yamazaki, T., J. Kondo, T. Watanabe, and T. Sato, 1992: A heat-balance model with a canopy of one or two layers and its application to field experiments. J. Appl. Meteor., 31, 86103, doi:10.1175/1520-0450(1992)031<0086:AHBMWA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yamazaki, T., H. Yabuki, Y. Ishii, T. Ohta, and T. Ohata, 2004: Water and energy exchanges at forests and a grassland in eastern Siberia evaluated using a one-dimensional land surface model. J. Hydrometeor., 5, 504515, doi:10.1175/1525-7541(2004)005<0504:WAEEAF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yamazaki, T., T. Ohta, R. Suzuki, and T. Ohata, 2007: Flux variation in a Siberian taiga forest near Yakutsk estimated by a one-dimensional model with routine data, 1986-2000. Hydrol. Processes, 21, 20092015, doi:10.1002/hyp.6708.

    • Search Google Scholar
    • Export Citation
  • Yoshida, R., M. Sawada, T. Yamazaki, and T. Iwasaki, 2011: Verification of nonhydrostatic-model simulated radiation using in situ observational data in Miyagi, Yamagata, and north part of Fukushima prefectures (in Japanese). Tenki, 31, 599605.

    • Search Google Scholar
    • Export Citation
  • Yoshida, R., T. Iizumi, and M. Nishimori, 2012: Inter-model differences in the relationships between downward shortwave radiation and air temperatures derived from dynamical and statistical downscaling models. J. Meteor. Soc. Japan, 90B, 7582, doi:10.2151/jmsj.2012-B06.

    • Search Google Scholar
    • Export Citation
  • Yukimoto, S., and Coauthors, 2012: A new global climate model of the Meteorological Research Institute: MRI-CGCM3—Model description and basic performance. J. Meteor. Soc. Japan, 90A, 2364, doi:10.2151/jmsj.2012-A02.

    • Search Google Scholar
    • Export Citation
  • Zhu, Y.-Y., H. Fang, Y.-Y. Wang, J. X. Fan, S.-S. Yang, T. W. Mew, and C. C. Mundt, 2005: Panicle blast and canopy moisture in rice cultivar mixtures. Ecol. Epidemiol., 95, 433438, doi:10.1094/PHYTO-95-0433.

    • Search Google Scholar
    • Export Citation
Save
  • Bonman, J. M., 1992: Durable resistance to rice blast disease—Environmental influences. Euphytica, 63, 115123, doi:10.1007/BF00023917.

    • Search Google Scholar
    • Export Citation
  • Bregaglio, S., M. Donatelli, R. Confalonieri, M. Acutis, and S. Orlandini, 2011: Multi metric evaluation of leaf wetness models for large-area application of plant disease models. Agric. For. Meteor., 151, 11631172, doi:10.1016/j.agrformet.2011.04.003.

    • Search Google Scholar
    • Export Citation
  • Bregaglio, S., M. Donatelli, and R. Confalonieri, 2013: Fungal infections of rice, wheat, and grape in Europe in 2030–2050. Agron. Sustainable Dev., 33, 767776, doi:10.1007/s13593-013-0149-6.

    • Search Google Scholar
    • Export Citation
  • Buser, C. M., H. R. Künsch, D. Lüthi, M. Wild, and C. Schär, 2009: Bayesian multi-model projection of climate: Bias assumption and interannual variability. Climate Dyn., 33, 849868, doi:10.1007/s00382-009-0588-6.

    • Search Google Scholar
    • Export Citation
  • Challinor, A. J., E. S. Simelton, E. D. G. Fraser, D. Hemming, and M. Collins, 2010: Increased crop failure due to climate change: Assessing adaptation options using models and socio-economic data for wheat in China. Environ. Res. Lett., 5, 034012, doi:10.1088/1748-9326/5/3/034012.

    • Search Google Scholar
    • Export Citation
  • Chen, C., G.-S. Zhou, and L. Zhou, 2014: Impacts of climate change on rice yield in China from 1961 to 2010 based on provincial data. J. Integr. Agric., 13, 15551564, doi:10.1016/S2095-3119(14)60816-9.

    • Search Google Scholar
    • Export Citation
  • Dunne, T., and L. N. Leopold, 1978: Water in Environmental Planning. W. H. Freeman and Company, 818 pp.

  • Earth System Grid Federation, 2014: ESGF portal. Program for Climate Model Diagnosis and Intercomparison, accessed 25 Aug 2014. [Available online at http://pcmdi9.llnl.gov/esgf-web-fe/.]

  • Garratt, J. R., and M. Segal, 1988: On the contribution of atmospheric moisture to dew formation. Bound.-Layer Meteor., 45, 209236, doi:10.1007/BF01066671.

    • Search Google Scholar
    • Export Citation
  • Gash, J. H. C., C. R. Lloyd, and G. Lachaud, 1995: Estimating sparse forest rainfall interception with an analytical model. J. Hydrol., 170, 7986, doi:10.1016/0022-1694(95)02697-N.

    • Search Google Scholar
    • Export Citation
  • Gautam, H. R., M. L. Bhardwaj, and R. Kumar, 2013: Climate change and its impact on plant diseases. Curr. Sci., 105, 16851691.

  • Gent, P. R., and Coauthors, 2011: The Community Climate System Model version 4. J. Climate, 24, 49734991, doi:10.1175/2011JCLI4083.1.

    • Search Google Scholar
    • Export Citation
  • Ghini, R., E. Hamada, F. Angelotti, L. B. Costa, and W. Bettiol, 2012: Research approaches, adaptation strategies, and knowledge gaps concerning the impacts of climate change on plant diseases. Trop. Plant Pathol., 37 (1), 524.

    • Search Google Scholar
    • Export Citation
  • Greer, C. A., and R. K. Webster, 2001: Occurrence, distribution, epidemiology, cultivar reaction, and management of rice blast disease in California. Plant Dis., 85, 10961102, doi:10.1094/PDIS.2001.85.10.1096.

    • Search Google Scholar
    • Export Citation
  • Hayashi, T., and Y. Koshimizu, 1988: Computer program BLASTAM for forecasting occurrence of rice leaf blast (in Japanese). Bull. Tohoku Natl. Agric. Exp. Stn., 78, 123138.

    • Search Google Scholar
    • Export Citation
  • Iizumi, T., M. Nishimori, Y. Ishigooka, and M. Yokozawa, 2010: Introduction to climate change scenario derived by statistical downscaling (in Japanese with English abstract). J. Agric. Meteor., 66 (2), 131143, doi:10.2480/agrmet.66.2.5.

    • Search Google Scholar
    • Export Citation
  • Iizumi, T., and Coauthors, 2012: Future change of daily precipitation indices in Japan: A stochastic weather generator-based bootstrap approach to provide probabilistic climate information. J. Geophys. Res., 117, D11114, doi:10.1029/2011JD017197.

    • Search Google Scholar
    • Export Citation
  • Ishizaki, N. N., and Coauthors, 2012: Improved performance of simulated Japanese climate with a multi-model ensemble. J. Meteor. Soc. Japan, 90, 235254, doi:10.2151/jmsj.2012-206.

    • Search Google Scholar
    • Export Citation
  • Kain, J. S., 2004: The Kain–Fritch convective parameterization: An update. J. Appl. Meteor., 43, 170181, doi:10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kawase, H., T. Yoshikane, M. Hara, B. Ailikun, F. Kimura, T. Yasunari, T. Inoue, and H. Ueda, 2009: Intermodel variability of future changes in the Baiu rainband estimated by the pseudo global warming downscaling method. J. Geophys. Res., 114, D24110, doi:10.1029/2009JD011803.

    • Search Google Scholar
    • Export Citation
  • Kerkhoff, C., H. Künsch, and C. Schär, 2014: Assessment of bias assumptions for climate models. J. Climate, 27, 67996818, doi:10.1175/JCLI-D-13-00716.1.

    • Search Google Scholar
    • Export Citation
  • Kim, K. S., S. E. Taylor, M. L. Gleason, and K. J. Koehler, 2002: Model to enhance site-specific estimation of leaf wetness duration. Plant Dis., 86, 179185, doi:10.1094/PDIS.2002.86.2.179.

    • Search Google Scholar
    • Export Citation
  • Kim, K. S., M. L. Gleason, and S. E. Taylor, 2006: Forecasting site-specific leaf wetness duration for input to disease-warning systems. Plant Dis., 90, 650656, doi:10.1094/PD-90-0650.

    • Search Google Scholar
    • Export Citation
  • Kobayashi, T., K. Ishiguro, T. Nakajima, H. Y. Kim, M. Okada, and K. Kobayashi, 2006: Effects of elevated atmospheric CO2 concentration on the infection of rice blast and sheath blight. Ecol. Epidemiol., 96, 425431.

    • Search Google Scholar
    • Export Citation
  • Kondo, J., T. Watanabe, M. Nakazono, and M. Ishii, 1992: Estimation of forest rainfall interception (in Japanese). Tenki, 39, 159167.

    • Search Google Scholar
    • Export Citation
  • Koshimizu, Y., 1988: A forecasting method for occurrence of rice leaf blast with AMeDAS data (in Japanese). Bull. Tohoku Natl. Agric. Exp. Stn., 78, 67121.

    • Search Google Scholar
    • Export Citation
  • Launay, M., J. Caubel, G. Bourgeois, F. Huard, I. G. Cortazar-Atauri, M.-O. Bancal, and N. Brisson, 2014: Climatic indicators for crop infection risk: Application to climate change impacts on five major foliar fungal diseases in northern France. Agric. Ecosyst. Environ., 197, 147158, doi:10.1016/j.agee.2014.07.020.

    • Search Google Scholar
    • Export Citation
  • Lee, J.-Y., and B. Wang, 2014: Future change of global monsoon in the CMIP5. Climate Dyn., 42, 101119, doi:10.1007/s00382-012-1564-0.

    • Search Google Scholar
    • Export Citation
  • Luo, W., and J. Goudriaan, 1999: Effects of altering water temperature on leaf wetness in paddy rice crops. Agric. For. Meteor., 97, 3342, doi:10.1016/S0168-1923(99)00046-5.

    • Search Google Scholar
    • Export Citation
  • Luo, Y., P. S. Teng, N. G. Febellar, and D. O. TeBeest, 1998: The effects of global temperature change on rice leaf blast epidemics: A simulation study in three agroecological zones. Agric. Ecosyst. Environ., 68, 187196, doi:10.1016/S0167-8809(97)00082-0.

    • Search Google Scholar
    • Export Citation
  • Magarey, R. D., J. M. Russo, and R. C. Seem, 2006: Simulation of surface wetness with a water budget and energy balance approach. Agric. For. Meteor., 139, 373381, doi:10.1016/j.agrformet.2006.08.016.

    • Search Google Scholar
    • Export Citation
  • Mashonjowa, E., F. Ronsse, M. Mubvuma, J. R. Milford, and J. G. Pieters, 2013: Estimation of leaf wetness duration for greenhouse roses using a dynamic greenhouse climate model in Zimbabwe. Comput. Electron. Agric., 95, 7081, doi:10.1016/j.compag.2013.04.007.

    • Search Google Scholar
    • Export Citation
  • Ministry of Agriculture, Forestry and Fisheries, 2014: Crop statistics (in Japanese). Accessed 12 May 2014. [Available online at http://www.maff.go.jp/j/tokei/kouhyou/sakumotu/sakkyou_kome/index.html#r.]

  • Mizuta, R., O. Arakawa, T. Ose, S. Kusunoki, H. Endo, and A. Kitoh, 2014: Classification of CMIP5 future climate responses by the tropical sea surface temperature changes. SOLA,10, 167–171, doi:10.2151/sola.2014-035.

  • Monteith, J. L., 1981: Evaporation and surface temperature. Quart. J. Roy. Meteor. Soc., 107, 127, doi:10.1002/qj.49710745102.

  • Nakanishi, M., and H. Niino, 2004: An improved Mellor–Yamada level-3 model with condensation physics: Its design and verification. Bound.-Layer Meteor., 112, 131, doi:10.1023/B:BOUN.0000020164.04146.98.

    • Search Google Scholar
    • Export Citation
  • Ogata, T., H. Ueda, T. Inoue, M. Hayashi, A. Yoshida, S. Watanabe, M. Kira, M. Ooshiro, and A. Kumai, 2014: Projected future changes in the Asian Monsoon: A comparison of CMIP3 and CMIP5 model results. J. Meteor. Soc. Japan, 92, 207225, doi:10.2151/jmsj.2014-302.

    • Search Google Scholar
    • Export Citation
  • Oh’izumi, M., I. Takayabu, and N. Ishizaki, 2013: Reproducibility of snow by a non-hydrostatic regional climate model (NHRCM) nested in JRA-25. J. Meteor. Soc. Japan, 91, 229238, doi:10.2151/jmsj.2013-209.

    • Search Google Scholar
    • Export Citation
  • Pedro, M. J., and T. J. Gillespie, 1981a: Estimating dew formation. I. Utilizing micrometeorological data. Agric. Meteor., 25, 283296, doi:10.1016/0002-1571(81)90081-9.

    • Search Google Scholar
    • Export Citation
  • Pedro, M. J., and T. J. Gillespie, 1981b: Estimating dew formation. II. Utilizing standard weather station data. Agric. Meteor., 25, 297310, doi:10.1016/0002-1571(81)90082-0.

    • Search Google Scholar
    • Export Citation
  • Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K. Averyt, M. Tignor, and H. L. Miller Jr., Eds., 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

  • Suzuki-Parker, A., I. Takayabu, R. Mizuta, H. Kusaka, K. Dairaku, S. Ham, S. A. Adachi, and N. N. Ishizaki, 2014: An overview of the SOUSEI multi GCM/RCM dynamical downscaling ensemble. Proc. 21st Century Challenges in Regional Climate Modelling, Lund, Sweden, International BALTEX Secretariat, 374–375. [Available online at http://www.baltex-research.eu/RCM2014/Material/IBESPS_No.3_low.pdf.]

  • Valente, F., J. S. David, and J. H. C. Gash, 1997: Modelling interception loss for two sparse eucalypt and pine forests in central Portugal using reformulated Rutter and Gash analytical models. J. Hydrol., 190, 141162, doi:10.1016/S0022-1694(96)03066-1.

    • Search Google Scholar
    • Export Citation
  • Wang, Z.-X., M. Yano, U. Yamanouchi, M. Iwamoto, L. Monna, H. Hayasaka, Y. Katayose, and T. Sasaki, 1999: The Pib gene for rice blast resistance belongs to the nucleotide binding and leucine-rich repeat class of plant disease resistance genes. Plant J., 19, 5564, doi:10.1046/j.1365-313X.1999.00498.x.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., and Coauthors, 2010: Improved climate simulation by MIROC5: Mean states, variability, and climate sensitivity. J. Climate, 23, 63126335, doi:10.1175/2010JCLI3679.1.

    • Search Google Scholar
    • Export Citation
  • Wichink Kruit, R. J., A. F. G. Jacobs, and A. A. M. Holtslag, 2008: Measurements and estimates of leaf wetness over agricultural grassland for dry deposition modeling of trace gases. Atmos. Environ., 42, 53045316, doi:10.1016/j.atmosenv.2008.02.061.

    • Search Google Scholar
    • Export Citation
  • Xie, S.-P., C. Deser, G. A. Vecchi, J. Ma, H. Teng, and A. T. Wittenberg, 2010: Global warming pattern formation: Sea surface temperature and rainfall. J. Climate, 23, 966986, doi:10.1175/2009JCLI3329.1.

    • Search Google Scholar
    • Export Citation
  • Yamazaki, T., 2001: A one-dimensional land surface model adaptable to intensely cold regions and its applications in eastern Siberia. J. Meteor. Soc. Japan, 79, 11071118, doi:10.2151/jmsj.79.1107.

    • Search Google Scholar
    • Export Citation
  • Yamazaki, T., J. Kondo, T. Watanabe, and T. Sato, 1992: A heat-balance model with a canopy of one or two layers and its application to field experiments. J. Appl. Meteor., 31, 86103, doi:10.1175/1520-0450(1992)031<0086:AHBMWA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yamazaki, T., H. Yabuki, Y. Ishii, T. Ohta, and T. Ohata, 2004: Water and energy exchanges at forests and a grassland in eastern Siberia evaluated using a one-dimensional land surface model. J. Hydrometeor., 5, 504515, doi:10.1175/1525-7541(2004)005<0504:WAEEAF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yamazaki, T., T. Ohta, R. Suzuki, and T. Ohata, 2007: Flux variation in a Siberian taiga forest near Yakutsk estimated by a one-dimensional model with routine data, 1986-2000. Hydrol. Processes, 21, 20092015, doi:10.1002/hyp.6708.

    • Search Google Scholar
    • Export Citation
  • Yoshida, R., M. Sawada, T. Yamazaki, and T. Iwasaki, 2011: Verification of nonhydrostatic-model simulated radiation using in situ observational data in Miyagi, Yamagata, and north part of Fukushima prefectures (in Japanese). Tenki, 31, 599605.

    • Search Google Scholar
    • Export Citation
  • Yoshida, R., T. Iizumi, and M. Nishimori, 2012: Inter-model differences in the relationships between downward shortwave radiation and air temperatures derived from dynamical and statistical downscaling models. J. Meteor. Soc. Japan, 90B, 7582, doi:10.2151/jmsj.2012-B06.

    • Search Google Scholar
    • Export Citation
  • Yukimoto, S., and Coauthors, 2012: A new global climate model of the Meteorological Research Institute: MRI-CGCM3—Model description and basic performance. J. Meteor. Soc. Japan, 90A, 2364, doi:10.2151/jmsj.2012-A02.

    • Search Google Scholar
    • Export Citation
  • Zhu, Y.-Y., H. Fang, Y.-Y. Wang, J. X. Fan, S.-S. Yang, T. W. Mew, and C. C. Mundt, 2005: Panicle blast and canopy moisture in rice cultivar mixtures. Ecol. Epidemiol., 95, 433438, doi:10.1094/PHYTO-95-0433.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Analysis area and calculation domain used in the dynamical downscaling. (a) The domain was 131 × 121 cells with a cell spacing of 20 km for NHRCM. (b) Leaf wetness observation sites: KWT (red dot), FRK (black dot), and KSM (green dot).

  • Fig. 2.

    Meteorological variables at the three observation sites in 2013: (a) surface air temperature, (b) downward shortwave radiation, (c) relative humidity, and (d) daily precipitation. See Fig. 1 for the locations of each of the three sites.

  • Fig. 3.

    Simulated surface air temperature changes based on (a) CCSM4, (b) MIROC5, and (c) MRI-CGCM3 and (d) the multimodel ensemble for future climate conditions (2081–2100, RCP4.5) relative to the period 1981–2000. (e)–(h) As in (a)–(d), but for daily precipitation changes. Each change is shown as the difference (surface air temperature) or ratio (daily precipitation) of the future climate to the present climate. The bottom-right values indicate the regional averages.

  • Fig. 4.

    As in Fig. 3, but for changes in the three precipitation indices: (a)–(d) MEA, (e)–(h) FRE, and (i)–(l) INT. The shading and values indicate the ratio of the precipitation index in the future climate (2081–99) to that in the present climate (1981–2000).

  • Fig. 5.

    As in Fig. 4, but for changes in leaf wetness variables: (a)–(d) LW, (e)–(h) number of wet events (NWE), and (i)–(l) averaged wet time per wet event (WT). A wet event is defined by the water amount on the leaf surface > 0 mm with a continuous duration ≥ 1 h.

  • Fig. 6.

    Scatterplot of the normalized leaf wetness and climate changes in the precipitation indices: (a)–(c) MEA, (d)–(f) FRE, and (g)–(i) INT. Each value was normalized as the ratio of the future climate minus the present climate relative to the present climate. Here R2, p, a, and b represent the determination coefficient, the p value, the slope, and the intercept of the regression line, respectively.

  • Fig. 7.

    As in Figs. 3e–g, but for the GCM-simulated changes in daily precipitation based on (a) CCSM4, (b) MIROC5, and (c) MRI-CGCM3. The changes are shown for the future climate (2081–2100) relative to the period 1981–2000. Each change is shown as the ratio of the future climate to the present climate.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 351 97 6
PDF Downloads 257 55 5