Climate Dynamics

, Volume 43, Issue 7, pp 1951–1963

Impact of land–sea breezes at different scales on the diurnal rainfall in Taiwan

Authors

    • Department of Earth SciencesNational Taiwan Normal University
  • Shih-Yu Wang
    • Utah Climate CenterUtah State University
    • Department of Plants, Soils, and ClimateUtah State University
Article

DOI: 10.1007/s00382-013-2018-z

Cite this article as:
Huang, W. & Wang, S. Clim Dyn (2014) 43: 1951. doi:10.1007/s00382-013-2018-z

Abstract

The formation mechanism of diurnal rainfall in Taiwan is commonly recognized as a result of local forcings involving solar thermal heating and island-scale land–sea breeze (LSB) interacting with orography. This study found that the diurnal variation of the large-scale circulation over the East Asia-Western North Pacific (EAWNP) modulates considerably the diurnal rainfall in Taiwan. It is shown that the interaction between the two LSB systems—the island-scale LSB and the large-scale LSB over EAWNP—facilitates the formation of the early morning rainfall in western Taiwan, afternoon rainfall in central Taiwan, and nighttime rainfall in eastern Taiwan. Moreover, the post-1998 strengthening of a shallow, low-level southerly wind belt along the coast of Southeast China appears to intensify the diurnal rainfall activity in Taiwan. These findings reveal the role of the large-scale LSB and its long-term variation in the modulation of local diurnal rainfall.

Keywords

Diurnal rainfallLand–sea breezeCirculation changeLow-level jet

1 Introduction

Located in the East Asia-Western North Pacific region (hereafter EAWNP; Fig. 1), the subtropical island of Taiwan exhibits large diurnal variations in surface wind and precipitation (Ramage 1952; Johnson and Bresch 1991; Dai and Deser 1999; Chen and Chen 2003; Dai et al. 2007; Wang and Chen 2008; Ruppert et al. 2013). Local solar heating and the interaction between Taiwan’s land–sea breeze (LSB) and orography are well-known mechanisms causing the late-afternoon rainfall maximum, which occurs around 5 p.m. local time (Li et al. 1997; Chen et al. 1999; Kishtawal and Krishnamurti 2001; Kerns et al. 2010). In addition to the late-afternoon rainfall, an increase in the early-morning rainfall appears around 5 a.m. local time (Yeh and Chen 1998; Kishtawal and Krishnamurti 2001; Chen et al. 2007; Wang and Huang 2009). This early-morning rainfall has been attributed to nocturnal drainage flows along the western slope of the mountains breaking the stable planetary boundary layer over the western plains of Taiwan (Chen et al. 1999; Wang and Huang 2009).
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig1_HTML.gif
Fig. 1

The geography and location of 21 conventional surface observation stations in Taiwan (dots). Stations are separated into three groups for estimating rainfall variations in western Taiwan (domain 1), central mountain range (hereafter, CMR) of Taiwan (domain2) and eastern Taiwan (domain 3) for use in Fig. 2. The scale of stippled orography in Taiwan is given in the right bottom panel

The large-scale circulation covering the EAWNP region also exhibits a marked diurnal variation (e.g. Yu et al. 2009). Huang et al. (2010) found that the LSB circulation over much of the East Asian coastal areas is coupled with the global-scale atmospheric pressure tide; this produces a planetary-scale LSB with a spatial scale of ~1,000 km. Later, Huang and Chan (2011, 2012) utilized a newer generation of global reanalysis, namely the Modern-Era Retrospective Analysis for Research and Applications (MERRA; Rienecker et al. 2011), to reveal further details of the large-scale LSB in EAWNP and the South China Sea. Huang and Chan (2011, 2012) showed that the interaction between the monsoonal southwesterlies and the continental LSB over East Asia contributes to the formation of morning convection in the South China Sea and afternoon convection in southern China. Because Taiwan is located within EAWNP’s diurnal flow regime, it is possible that the formation mechanism of diurnal rainfall in Taiwan also undergoes certain modulation by such a large-scale LSB. This aspect has not been documented and is examined herein.

It is anticipated that the much increased temporal and spatial resolutions of modern reanalyses like MERRA can provide an observational depiction of the interaction between the island-scale LSB (referred to as Taiwan-LSB) and the large-scale LSB (referred to as EAWNP-LSB), as well as their impact on the formation of local diurnal rainfall. Both the Taiwan-LSB and EAWNP-LSB exhibit a marked seasonal variation (Li et al. 2008; Huang et al. 2010; Kerns et al. 2010; Huang and Chan 2012). Here, our season of interest is May and June (MJ), a time period when the diurnal rainfall in Taiwan experiences a strong interaction between Taiwan-LSB and EAWNP-LSB (explained later). Analyses in this paper were performed using various sources of observational datasets introduced in Sect. 2. Results are shown in Sect. 3. A summary and conclusion are provided in Sect. 4.

2 Data sources

Observed precipitation is derived from (1) 21 conventional meteorological stations operated by Taiwan’s Central Weather Bureau as indicated in Fig. 1 and (2) 3-hourly TRMM (Tropical Rainfall Measuring Mission) 3B42 satellite precipitation (Simpson et al. 1996) beginning in 1998. The TRMM 3B42 dataset provides rain rate at the spatial resolution of 0.25° longitude × 0.25° latitude, comparable with the common spacing of meteorological station network in EAWNP (e.g. Zhou et al. 2008). Satellite observed rain rate has proven to be a good proxy for diurnal convection (e.g. Dai et al. 2007; Mao and Wu 2012). Kishtawal and Krishnamurti (2001) found that the TRMM 3B42 rain rate replicates the gauge-depicted diurnal convection features in Taiwan and provides additional information not observable at the surface, such as vertical profile or convection nearby surrounding oceans. Hong et al. (2005) also showed that the TRMM precipitation represents well the diurnal rainfall variability in EAWNP.

Meteorological data including wind fields, humidity, and vertical velocity are extracted from the 3-hourly MERRA reanalysis (Rienecker et al. 2011) at the spatial resolution of 0.667° longitude × 0.5° latitude. MERRA’s spatial and temporal resolutions represent a leap forward when compared to older reanalyses with the common 6-hourly, 2.5-deg resolutions. The analysis here covers the time period from 1998 to 2012 for the MJ months, when diurnal variation is becoming predominant in the region (Wang and Chen 2008). The following analyses refer to the local time in Taiwan which is universal time (UTC) + 8 h, e.g. 0800 h is 0000 UTC, and so forth.

3 Results

3.1 Rainfall characteristics

Based on the longitudinally stratified orographic features in Taiwan, we define three sub-regions consisting of western Taiwan with plains and relatively gradual slopes (domain 1 in Fig. 1), the central mountain range (CMR; domain 2), and eastern Taiwan with steep slopes (domain 3). The temporal evolution of the 3-hourly rainfall averaged from 21 selected surface stations (Fig. 2a) shows a bimodal signal, with an early-morning peak at 0800 h and a late-afternoon maximum at 1700 h. In each sub-region, western Taiwan shows the pronounced bimodal signal (Fig. 2b), while the CMR and eastern Taiwan exhibit a single late-afternoon peak of diurnal rainfall (Fig. 2c, d). However, the timing of the rainfall peak is skewed differently towards the afternoon hours of 1400–1700 h in the CMR and the evening hours of 1700–2000 h in eastern Taiwan. In Fig. 2 we overlay the TRMM rain rate averaged over Taiwan and from the three sub-regions. The general resemblance between the station and TRMM observations indicates that TRMM indeed replicates the timing and amplitude of diurnal rainfall across Taiwan; although the actual rainfall values are underestimated.
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig2_HTML.gif
Fig. 2

Time evolution of 3-hourly precipitation (P) estimated from rain gauge stations (histogram) and TRMM 3B42 (solid line with dots) for a the entire Taiwan, b the western Taiwan (domain 1 in Fig. 1), c the CMR of Taiwan (domain 2) and d the eastern Taiwan (domain 3). The rainfall scale for rain gauge observation is shown in the left and for the TRMM observation in the right. The lower x-axis is local time in Taiwan, which is universal time (UTC) +8 h

Next, the regional features of the TRMM precipitation over the EAWNP region are shown in Fig. 3, derived from the MJ climatology of 1998–2012. Rainfall first appears at 0200 h over the Taiwan Strait and develops along the western coast of Taiwan up to 0800 h (Fig. 3a–c). Morning rainfall is also observed along the south China coast (e.g. Aves and Johnson 2008). Beginning early afternoon (1400 h), heavy rainfall quickly develops over Taiwan through 1700 h (Fig. 3d–f). At 2000 h, rainfall dissipates over the island and yet, a narrow rainband develops over ocean along the eastern coast of Taiwan into midnight (Fig. 3g, h). This nighttime development of coastal rainband was observed by Yu and Lin (2008) and Alpers et al. (2010).
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig3_HTML.gif
Fig. 3

Horizontal distributions of the TRMM 3B42 rain rate averaged during 1998–2012 May and June at a 0200, b 0500, c 0800, d 1100, e 1400, f 1700, g 2000 and h 2300 h (local time in Taiwan). The color scale of ah is given on top of e

3.2 LSB scale interaction

To provide a quantitative measure of the diurnal wind variation in Taiwan, we compute the surface wind divergence (Ds) from stations following the method used in Chen et al. (1999):
$$D_{s} = \frac{1}{A}\int\limits_{L} {V_{n} dl} ,$$
(1)
where A, Vn and dl indicate respectively the area of Taiwan encircled by the circumference connecting the selected stations (Fig. 4b), the surface wind velocity normal to the circumference and the distance increment along the circumference. It is expected that solar radiative heating drives the temporal evolution of the island-averaged Ds and its maximum convergence (i.e. −Ds > 0) at 1400 h, following the maximum surface temperature; this is confirmed in Fig. 4a. Next, DS is derived from MERRA’s 10 m winds over the domain of 120°–122°E, 22°–25.5°N (surrounding Taiwan). As shown in Fig. 4a, MERRA’s DS is in good agreement with that derived from surface stations. This resemblance of MERRA with the station-observed LSB is remarkable and thus builds confidence in our subsequent analysis for the scale interaction of LSBs.
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig4_HTML.gif
Fig. 4

a Time evolution of 3-hourly surface wind convergence (-Ds) estimated from selected surface stations (solid line) and the MERRA reanalysis (dotted line). In a, time evolution of 3-hourly surface temperature averaged from all 21 stations [black + white dots marked in b] is overlaid as histogram. The methodology used for the estimation of (-Ds) from selected surface stations (black dots with solid line linked) is explained in b. The scale of stippled orography in b is given in the right

The diurnal variation of the surface circulation over EAWNP is shown in Fig. 5 and applied with the first harmonics (S1), using the Fourier analysis in order to isolate the diurnal component [following Huang et al. (2010) who found that the S1 component of surface winds explains 90 % of the diurnal variance in EAWNP]. Between 0800 h (Fig. 5c) and 2000 h (Fig. 5g), the diurnal variation of lower-tropospheric winds reveals a reversal of the large-scale LSB over the EAWNP region (Huang et al. 2010). At 0200 h (Fig. 5a), land breezes from western Taiwan and southeastern China collide and form an apparent convergence line along the Taiwan Strait and coastal southern China. By 0800 h the local land breeze in western Taiwan has weakened while the land breeze in southeastern China persists (Fig. 5c), which later dissipates in the early afternoon. Such a continental land breeze is part of EAWNP-LSB as was documented in Huang et al. (2010). From 0200 to 1100 h, the interaction of land breezes between Taiwan and southeastern China migrates eastward and displaces the early morning convergence over the Taiwan Strait and then over Taiwan. The initiation of the relatively strong island convergence at 1100 h occurs soon after this regional-scale convergence zone propagates through. The propagating convergence is indicated by the white dashed lines in Fig. 5, from which the location are based upon the maximum center of S1 of the 925-hPa vertical velocity (see next).
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig5_HTML.gif
Fig. 5

The diurnal harmonic (S1) of 925 hPa vertical motion (-ω) superimposed with 10 m wind field. The color scale of (−ω) is given on top of e. The topography with altitude at 500, 1,000, 2,000 and 3,000 m are marked by the green lines. The white dashed line indicates the movement of the regional convergence zone that located based upon the maximum center of S1(−ω) at 925 hPa

3.3 Impact on diurnal rainfall

The consequence from such a scale interaction between the local and regional LSBs can be visualized by the 925-hPa vertical motion superimposed with Fig. 5. Overall, areas with surface convergent anomalies are associated with ascending motion with the strength corresponding to the magnitude of convergence; areas of divergence are association with descending motion. Examining the early-morning rainfall in Taiwan, Chen et al. (1999) suggested that the nocturnal inversion along Taiwan’s western plains is broken by downslope flow of cooled air, triggering shallow early morning convection. Using MERRA and based upon the regional perspective, we find that the early morning convergence initiated in the Taiwan Strait also plays a role as it propagates over coastal western Taiwan, forming the morning convergence there and the subsequent morning rainfall (Fig. 2b).

The formation mechanism of Taiwan’s afternoon convection is rather straightforward, i.e. sea breeze interacting with terrain in conjunction with heated slopes creating thermal lifting (Wang and Chen 2008). However, additional effect of EAWNP-LSB on Taiwan’s afternoon convection can be observed from Fig. 5c–e, in which the development of the local convergence in Taiwan is coupled with the passage of the regional surface convergence zone. Soon after the regional surface convergence zone propagates over Taiwan, a relatively strong island convergence is initiated at 1100 h (Fig. 5d) and persists through 1700 h under the strong local thermal lifting. As the surface convergence zone moves away from Taiwan (Fig. 5e, f), easterly winds develop over the ocean as part of the sea breeze of EAWNP-LSB (Huang et al. 2010; Huang et al. 2013) and encounter the land breeze in eastern Taiwan; this forms a subsequent narrow band of convergence offshore eastern Taiwan (~0200 h, Fig. 5a).

The vertical cross sections of S1 of zonal wind and vertical velocity (Fig. 6) depict further details of the LSBs’ scale interaction. Upward motion near Taiwan appears to be coupled with the propagating low-level convergence (connected with red lines in the figure). The pair moves across the Taiwan Strait (~0500 h), Taiwan island (1100 h), and then integrates into strong daytime upward motion over the island (~1700 h). The evening convergence zone in eastern Taiwan, formed by the converging local land breeze and the remnant of the large-scale sea breeze (2000–0500 h), also exhibits an eastward propagation. The propagating upward motion later integrates into the large-scale ascent from 0800 h onwards; this is indicated by the red line over the area east of Taiwan. This convergence band echoes the radar observation of Alpers et al. (2007) and Yu and Lin (2008), who showed that nighttime convective lines generally form offshore eastern Taiwan between 2000 and 2200 h. Using MERRA, we further realize that the large-scale sea breeze, which remains pronounced at 2000 h (Fig. 6g), encounters the initial stage of the shallow offshore winds in eastern Taiwan and produces the elongated cloud band about 40–80 km offshore (Alpers et al. 2007). As a result, the diurnal rainfall variation in eastern Taiwan is skewed toward 2000 h in comparison with other parts of Taiwan (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig6_HTML.gif
Fig. 6

The 3-hourly vertical cross-section of the diurnal harmonic (S1) of zonal wind (u) and vertical motion (−ω) averaged between 23° and 24°N. The color scale of S1(−ω) is given on top of a. The red lines are plotted to indicate the propagation of the convergence/upward motion zones

3.4 Synoptic modulations

The frequency and intensity of Taiwan’s diurnal rainfall activity are modulated by the synoptic conditions, such as the quasi-biweekly (QBW) mode observed by Wang et al. (2013a). When the cyclonic circulation of the QBW mode develops to the west of Taiwan followed by an anticyclonic circulation to the east, the diurnal convection can intensify and the intensification can persist for several days. This synoptic setting reflects increased monsoonal southwesterly winds and moistened lower troposphere while the enhanced subtropical anticyclone favors fair weather conditions with increased solar radiative heating. To examine the impact of EAWNP-LSB under such synoptic conditions, we categorize the active (inactive) diurnal rainfall days in Taiwan, defined as the daily amplitude of diurnal rainfall is greater (smaller) than 0.2 mm h−1. In this composite analysis, dates with the influences from tropical cyclones and fronts are eliminated by following the criteria used in Wang et al. (2013a).

As shown in Fig. 7, the anomalous surface flows between the active and inactive diurnal rainfall days reveal a low center over coastal southern China associated with enhanced southwesterly flows towards Taiwan; the flows then curve into southeasterlies over eastern and northern Taiwan. The circulation anomalies shown in Fig. 7 lead to an enhancement of moisture transported from the South China Sea (e.g. Chen et al. 2013), which in turn increases the moisture supply for the initiation and maintenance of diurnal rainfall in Taiwan. Because this analysis is for the MJ season, the anomalous circulation between the active and inactive diurnal rainfall days could also enhance synoptic/meiyu fronts (Geng and Yamada 2007; Sun and Zhang 2012), as is revealed by the rainband structure visible throughout the day even though frontal events have been eliminated in the composite analysis. Such synoptic conditions apparently enhance the early morning rainfall in western Taiwan as well (Fig. 7c).
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig7_HTML.gif
Fig. 7

Composite differences of Vs (10 m wind field) and P (precipitation) between the active and inactive diurnal rainfall days in Taiwan. The active and inactive diurnal rainfall days in Taiwan are defined as the date with the amplitude of diurnal rainfall in Taiwan ≥ (<) 0.2 mm h−1, respectively. The color scale of P is given on top of e. Only the values of Δ(Vs, P) exceeding the 90 % confidence level are shown

3.5 Change in the recent decade

Given the connection of EAWNP-LSB with Taiwan-LSB, as well as the changing circulation/sea surface temperature pattern observed in the region (e.g. Wang et al. 2013b), the long-term change of diurnal rainfall in Taiwan is examined here based on the available TRMM 3B42 data during the 1998–2012 MJ seasons. The circulation change is shown in Fig. 8 through the comparison between two eras: 2006–2012 and 1998–2004. Over the entire Taiwan (Fig. 8a) the daily rainfall has increased by about 25 % while the morning (0800 h) and afternoon (1700 h) rainfall has increased more than that at other hours, suggesting an amplification of the diurnal cycle. This overall change in diurnal rainfall consists of sharp increases in the morning rainfall over western Taiwan (Fig. 8b), afternoon rainfall in central Taiwan (Fig. 8c), and evening rainfall in eastern Taiwan (Fig. 8d).
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig8_HTML.gif
Fig. 8

Similar to Fig. 2 but for the 3-hourly precipitation estimated from TRMM 3B42 averaged during 1998–2004 (histogram) and 2006–2012 (solid line with dots)

The surface wind and regional precipitation changes between the two eras are subsequently examined to inspect the large-scale circulation change associated with the diurnal rainfall change; this is shown in Fig. 9. The first impression from such long-term change of surface wind field is its resemblance with the synoptic setting for active diurnal activity in Taiwan (Fig. 7). There is also an apparent narrow band of southwesterly winds formed along southeastern China, accompanying the rainband that extends over to southern Japan. Such wind and precipitation anomalies persist throughout the day but also reveal discernible diurnal fluctuations. The corresponding anomalies of the diurnal wind and precipitation (not shown) depict a pattern similar to that of Fig. 5, suggesting an intensification in both Taiwan-LSB and EAWNP-LSB and the precipitation response.
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig9_HTML.gif
Fig. 9

Similar to Fig. 7 but for the composite differences of (Vs, P) between the 2006–2012 and 1998–2004 eras. The northwest–southeast orange line added at 0800 h/0000 UTC indicates the vertical cross-section in Fig. 10. Only the values of Δ(Vs, P) exceeding the 90 % confidence level are shown

The anomalous winds in Fig. 9 signify a possible intensification of the low-level jet (LLJ), a regional circulation feature critical to rainfall formation in Taiwan during the MJ season (Chen 1983, 1993; Chen et al. 2008). The LLJ in this region brings the monsoon moisture from the South China Sea to the Taiwan Strait and, when encountering the mountain range in Taiwan, forms a localized jet core over the western slopes fueling local convection. In Fig. 10 we show the vertical sections of (a) the meridional wind and (b) the meridional moisture flux differences across Taiwan (indicated by the orange line in Fig. 9) at 0800 h between the two eras. The LLJ core is under 800 hPa to the west of Taiwan (Fig. 10a) while the southerly component of both wind and moisture flux is confined below the top of Taiwan’s mountain range (Fig. 10b). Also noteworthy is the width of the LLJ intensification that engulfs Taiwan. Wang et al. (2013a) have shown that increased southerly or southwesterly winds originated from the South China Sea enhance conditional instability due to lower tropospheric moistening, while the increased upslope winds over the mountains (or coastal convergence owing to land breeze) provide lift, hence favoring diurnal rainfall in Taiwan. Therefore, the recent increase of diurnal rainfall variability in Taiwan (cf. Fig. 8) appears to be dynamically and thermodynamically supported by the increased, moistened low-level southerly winds. The causes of the LLJ intensification in the recent years would require more sophisticated analyses for careful explanation and attribution.
https://static-content.springer.com/image/art%3A10.1007%2Fs00382-013-2018-z/MediaObjects/382_2013_2018_Fig10_HTML.gif
Fig. 10

a Vertical cross-section of the composite differences in a meridional wind v and b meridional moisture flux vq at 0800 h/0000 UTC between the 2006–2012 and 1998–2004 eras. The cross-section is made along the northwest–southeast line indicated in Fig. 9. The contour interval in a and b respectively is ms−1 and 10−2 ms−1·g kg−1. Only the differences exceeding the 90 % confidence level are shaded

4 Summary and conclusion

The impact of the large-scale diurnal circulations on Taiwan’s diurnal rainfall formation was examined for the MJ seasons during 1998–2012 using satellite data and the high-resolution MERRA reanalysis. The analyses showed that the diurnal rainfall in Taiwan consists of three regimes: a noticeable early-morning rainfall in western Taiwan, a predominant afternoon rainfall in central Taiwan, and a marked nighttime rainfall in eastern Taiwan. Analyses conducted by MERRA suggested that such sub-island differences are caused by the interaction between the two land–sea breeze systems: Taiwan-LSB (i.e. island-scale LSB) and EAWNP-LSB (i.e. large-scale LSB). In the early morning, easterly winds from the land breeze of Taiwan-LSB meet the westerly winds from the large-scale land breeze of EAWNP-LSB; their interaction initiates surface convergence in the Taiwan Strait. When coupled with the large-scale land breezes that migrate eastward across the Taiwan Strait, the shallow convergence facilitates the formation of early morning rainfall in western Taiwan. Later in the day, the late-afternoon rainfall in central Taiwan coincides with the passage of the large-scale surface convergence zone, which contributes to the local thermal forcing in the initiation of diurnal convection. At the nighttime, the shallow outflows in eastern Taiwan collide with easterly winds from the remnant of the large-scale sea breeze of EAWNP-LSB, triggering the formation of nighttime rainfall over the coastal ocean.

Additional synoptic analyses post-1998 indicated an intensification of a narrow band of southwesterly flows coming from the South China Sea towards Taiwan. Cross-sections of such circulation changes revealed a wind pattern resembling the LLJ—i.e. shallow southerly winds with a near-ground, high wind speed jet core to the immediate west of Taiwan. The increased LLJ provides favorable synoptic conditions of lift and instability for diurnal convection to develop over Taiwan. These findings may provide guidance to weather forecasting, and pointed out the need for further studies using sophisticated climate models in projecting future changes of Taiwan’s frequent diurnal rainfall.

Acknowledgments

The authors thank anonymous reviewers for their comments and suggestions which greatly improved the manuscript. This research was supported by the National Science Council of Taiwan under NSC 101-2119-M-003-006-MY2 and NSC 102-2621-M-492-001. SYW was supported by Grant MOTC-CWB-101-M-15, Grant NNX13AC37G, and the Utah Agricultural Experiment Station, Utah State University as Journal Paper # 8464.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013