Midday measurements of leaf water potential and stomatal conductance are highly correlated with daily water use of Thompson Seedless grapevines
- First Online:
- Cite this article as:
- Williams, L.E., Baeza, P. & Vaughn, P. Irrig Sci (2012) 30: 201. doi:10.1007/s00271-011-0276-2
- 2.1k Downloads
A study was conducted to determine the relationship between midday measurements of vine water status and daily water use of grapevines measured with a weighing lysimeter. Water applications to the vines were terminated on August 24th for 9 days and again on September 14th for 22 days. Daily water use of the vines in the lysimeter (ETLYS) was approximately 40 L vine−1 (5.3 mm) prior to turning the pump off, and it decreased to 22.3 L vine−1 by September 2nd. Pre-dawn leaf water potential (ΨPD) and midday Ψl on August 24th were −0.075 and −0.76 MPa, respectively, with midday Ψl decreasing to −1.28 MPa on September 2nd. Leaf gs decreased from ~500 to ~200 mmol m−2 s−1 during the two dry-down periods. Midday measurements of gs and Ψl were significantly correlated with one another (r = 0.96) and both with ETLYS/ETo (r = ~0.9). The decreases in Ψl, gs, and ETLYS/ETo in this study were also a linear function of the decrease in volumetric soil water content. The results indicate that even modest water stress can greatly reduce grapevine water use and that short-term measures of vine water status taken at midday are a reflection of daily grapevine water use.
Irrigated agriculture needs to become more water use efficient due to increasing demands and when annual rainfall within a crop production area is below normal (Larus 2004; Morison et al. 2008). This is especially true in California where the competition among environmental, agricultural and urban demands for limited water supplies continues to increase. Grapevines grown at most locations in California are irrigated at least some time during the growing season due to the fact that the majority of rainfall will occur during the dormant portion of the growing season and that grapevine water use will exceed the amount of water held in the soil reservoir (Williams and Matthews 1990). Seasonal evapotranspiration (ETc) of young to mature grapevines grown in the San Joaquin Valley can range from 200 to 800 mm, while estimated ETc of mature wine grape vineyards at different locations in California may vary from 450 to 800 mm per growing season (Williams and Baeza 2007; Williams et al. 2003a, b; Williams and Ayars 2005a, b). Differences in ETc among and within the different grape commodities are a function of trellis type and row spacing (Williams and Ayars 2005b).
Fereres and Soriano (2007) have suggested that the potential for water conservation in tree and vine crops may be greater than those of field crops since they are more highly coupled to the atmosphere. However, they also pointed out that the use of deficit irrigation in orchards and vineyards will require close monitoring of soil and/or plant water status so as to minimize risk. There are numerous tools one can use to monitor vine water status or assist in scheduling vineyard irrigations (Cifre et al. 2005); Jones 2004).
Currently, the most widely used technique to monitor vine water status in California’s commercial vineyards is midday Ψl and to a lesser extent midday Ψstem (L.E. Williams, personal observation). Recently, porometers are being used in the north coast wine production regions (Sonoma and Napa counties) of California by both consultants and growers to aid in irrigation management (R. H. Smith, personal communication). Soil water availability also is being monitored with neutron probes by a few crop consultants, while some grape growers may use tensiometers or electrical resistance type sensors. For the aforementioned techniques to be useful in an irrigation management program, the determination as to whether they are actually correlated with grapevine water use would be beneficial.
A weighing lysimeter was installed in 1986 (Williams et al. 2003a) and used to measure water use of two Thompson Seedless grapevines to develop seasonal crop coefficients for the San Joaquin Valley of California (Williams et al. 2003a, b; Williams and Ayars 2005a, b). During August of 2005, the lysimeter pump was turned off for 9 days, irrigation resumed for 11 days and then the pump turned off a second time for 21 days. During this time frame, midday Ψl and gs and soil volumetric water content were measured frequently to determine whether daily grapevine water use during these periods was correlated with these short-term measures of vine/soil water status. Measurements were taken at midday since grapevine stomatal conductance and Ψl reach their maximum and minimum diurnal values, respectively, during this time frame (Grimes and Williams 1990; Williams and Baeza 2007). It was also felt that limiting these measurements to the two vines growing in the lysimeter would reduce potential plant to plant variation one may encounter in vineyards (van Leeuwen et al. 2006) and orchards (Naor et al. 2006) while accurately measuring water use of mature grapevines with a weighing lysimeter (Allen et al. 1998).
Materials and methods
A weighing lysimeter was installed at the University of California’s Kearney Agricultural Center located in the San Joaquin Valley of California (36°48′N, lat, 119°30′W, long.) in 1986. Two Vitis vinifera L. (cv. Thompson Seedless clone 2A) grapevine cuttings were planted in the lysimeter and in the surrounding 1.4-ha (168 m × 82 m) vineyard on 9 April 1987. The soil was a Hanford fine sandy loam (coarse-loamy, mixed, non-acid, thermic Typic Xerorthent). The trellis of the vines used in the study consisted of a 2.13-m wooden stake driven 0.45 m into the soil at each vine. A 0.6-m cross-arm was placed atop the stake and wires attached at either end of the cross-arms to support the vine’s fruiting canes. Row direction was approximately east/west. A detailed description of the lysimeter and calculations used to measure grapevine evapotranspiration (ETc) are given elsewhere (Williams et al. 2003a, b). Reference ET (ETo) data were obtained from a California Irrigation Management Information System (CIMIS) weather station located 2 km from the vineyard site. Variables measured and calculations used to determine hourly and daily ETo from CIMIS can be found in Synder and Pruitt (1992).
Prior to the initiation of the study, two layers of 30-gauge, clear PVC film (Goss Plastic Film Corp., Los Angeles, CA) were used to cover the soil surface of the lysimeter on August 22 (day of year [DOY] 234) and held in place with weights. The soil remained covered by the plastic throughout the study except on September 13th and 14th (DOYs 256 and 257). On those days, the plastic was rolled to the edges of the lysimeter, exposing the wetted soil surface. Water use of the grapevines in the lysimeter will be designated ETLYS to denote the fact that the soil surface was covered with plastic, so as to minimize soil evaporation (Steinberg et al. 1990).
The irrigation pump to the lysimeter was turned off on August 24 (DOY 236) at 1700 h. The pump was turned back on September 2 (DOY 245) at 1800 h. The next day, the lysimeter pump was run for five hours to supply additional water (~120 L vine−1) to the soil. Thereafter, the lysimeter pump was activated when the vines used the equivalent of 2 mm of water or 8 L vine−1. The lysimeter pump was turned off again on September 14 (DOY 257) at 1700 h and remained such until October 4 (DOY 278). The pump was activated for 1.5 h the next day and then allowed to irrigate as stated above.
Water potential measurements were conducted as described by Williams and Araujo (2002). Specifically, pre-dawn Ψ (ΨPD) measurements began at ≈0430 h and were finished prior to sunrise using a pressure chamber (Model 1000, PMS Instrument Co., Corvallis, Ore.). Midday measurements of leaf water potential (Ψl) generally were taken between 1230 and 1330 h, Pacific Daylight Time (PDT). Leaf blades for ΨPD and Ψl determinations were covered with a plastic bag, quickly sealed, and petioles then cut within 1–2 s. The time between leaf excision and chamber pressurization was generally <10–15 s. Leaves, chosen for ΨPD and Ψl were fully expanded and mature. At midday, Ψl was measured on leaves exposed to direct solar radiation located on the top or south sides of the canopy. Two to three leaves from each of the lysimeter vines were measured and used for data analyses. While the number of leaves removed from each vine in the lysimeter to measure Ψl would have been upwards of 60 per vine, the total number of leaves from primary shoots on a mature Thompson Seedless grapevine can be greater than 1,500 (Williams 1987) with a total leaf area in excess of 20 m2 vine−1 (Williams 1987; Williams et al. 2003b, 2010a). Therefore, the removal of these leaves would only have had a minimal effect, if any, on water use and/or water status of these vines.
Soil water content (SWC) within the lysimeter was monitored using the neutron back-scattering technique with a neutron moisture probe (Model 503 DR Hydroprobe moisture gauge: Boart Longyear, Martinez, California). Two access tubes were placed 0.5 m from each vine within the lysimeter (approximately 1.0 m between the two tubes beneath the drip line) and inserted to a depth of 1.5 m. Readings were taken at depths of 0.23, 0.46, 0.76, 1.07, and 1.37 m from the soil surface. Field capacity of this soil type was approximately 22.8 percent by volume (θv), while SWC at a soil moisture tension of −1.5 MPa was approximately 8.0 θv (Araujo et al. 1995).
A comparison of midday Ψl and gs between the two vines growing in the lysimeter and vines growing elsewhere in the vineyard were made in response to: (1) the termination of irrigation during the first dry-down period and (2) the continued irrigation of vines during both dry-down periods. The vineyard was divided into quadrants and irrigation terminated in 3, half-rows during the first dry-down period, beginning August 24 (DOY 236). The continuously irrigated vines were in the same half-row in which the lysimeter was located during both dry-down periods.
The irrigation pump for the rest of the vineyard, to include the row the lysimeter was located, was controlled by the lysimeter’s datalogger (Campbell Scientific 21X Micrologger, Logan, Utah). Whenever vines within the lysimeter used 2 mm of water, the vineyard pump was activated and an irrigation event took place with an equivalent amount of water; 8 L vine−1. The activation of a solenoid valve at the head of the row provided the appropriate amount of applied water. An in-line water meter upstream from the solenoid valve measured actual applied water amounts. It should be pointed out that the amount of water applied to these vines, once the lysimeter pump was turned off, slowly decreased and was less than that normally applied to vines at full ETc.
Data were analyzed via correlation analysis or regression analysis using linear, quadratic, cubic, or other terms. Values presented in the figures represent the means of measurements from the two lysimeter-grown vines for midday Ψl and gs versus ETLYS on the day measurements were taken. Measurements of Ψl and gs taken on vines growing outside the lysimeter represent the means of 6 individual replicates for Ψl and 9 for gs (two and three measurements per row for the vines outside the lysimeter for Ψl and gs, respectively). Differences in Ψl and gs between days or measurements made on vines in the lysimeter, the irrigated vines or the non-irrigated vines located elsewhere in the vineyard were analyzed via analysis of variance and means separated using Tukey’s test.
Comparisons of midday leaf water potential (Ψl—MPa) and stomatal conductance (gs—mmol m−2 s−1) for the lysimeter-grown vines and vines growing in the surrounding vineyard receiving no water during the first dry-down period (1st ddp)
Day of year
Vineyard vines (1st ddp)
No applied water vines
The significant correlations between midday Ψl and gs and midday Ψl (and gs) and the ETLYS/ETo ratio found in this study were similar to those of Mata et al. (1999). They found that the decrease in daily peach (Prunus persica L.) tree transpiration (measured with a weighing lysimeter) was linearly related to midday Ψstem during a dry-down period of 20 days (deficit irrigation followed by no-irrigation). The relationship between lysimeter grapevine Ψstem and ETLYS/ETo would probably have been similar to that of Mata et al. (1999) in this study if Ψstem had been measured as it and Ψl are highly correlated with one another in grapevines (Salón et al. 2005; Stevens et al. 1995; Williams and Araujo 2002; Williams and Trout 2005). Mata et al. (1999) also found that tree water use and Ψstem were highly correlated with soil water content, similar to that found in this study for Ψl and gs and soil water content. Williams and Trout (2005) previously had found that midday Ψstem and Ψl were highly correlated with soil water content when measured throughout the season in this same vineyard. Lastly, it has been demonstrated that the ETc/ETo ratio was linearly related to ΨPD of table grapes grown in Italy (Rana et al. 2004). A similar result was found in this study, albeit using only four values of ΨPD.
There are a few studies in which short-term measures of plant water status and daily or hourly whole plant water use, measured with sap flow sensors, have been compared in grapevines. Sousa et al. (2006) found that sap flow of grapevines measured at 1400 h was significantly correlated (r = 0.75) with Ψl measured at 1400 h but not with gs measured at the same time. They also found that sap flow and Ψl measured at 1400 h were significantly correlated with soil water content. In that study, mean Ψl measured at 1400 h was −1.17 and −1.45 MPa for the fully irrigated and non-irrigated treatments, respectively. Patakas et al. (2005) found that the ratio of the mean daily sap flow of their water stressed vines (SFTi) divided by the mean daily sap flow of their fully irrigated vines (SFTl) was highly correlated (R2 = 0.83) with midday measurements of Ψstem ranging from −0.35 to −1.15 MPa. No such relationship was found when the ratio was compared to measures of Ψl (midday Ψl averaged −1.45 MPa for all treatments in that study). Escalona et al. (2002) found that sap flow and short-term measures of leaf gs and transpiration of irrigated and non-irrigated, potted grapevines differed significantly, but Ψl did not differ between the two treatments.
While the Sousa et al. (2006) study somewhat agrees with results presented in this paper (relationship between midday Ψl and grapevine water use) those of Patakas et al. (2005) and Escalona et al. (2002) do not. One possible explanation for the differences between the two latter studies mentioned above and this study could be related to the method and/or technique in which Ψl was measured. It has been demonstrated that the failure to enclose the leaf blade in a plastic bag just prior to cutting the petiole can dramatically affect the values of Ψl one measures (Turner and Long 1980). For example, the senior author of this study found that midday Ψl of fully irrigated vines used in this study on a high evaporative day was 0.6 MPa lower for leaves that were not bagged versus bagged leaves (−1.3 vs. −0.7 MPa, respectively). In addition, the difference in Ψl values between bagged and non-bagged leaf blades is greater in rapidly transpiring leaves compared to those that are not (Turner and Long 1980; Williams and Araujo 2002). Leaf Ψ was measured in the Sousa et al. (2006) study by enclosing the leaf blade in a plastic bag prior to cutting the petiole (noted in their “Materials and methods” section). Patakas et al. (2005) only enclosed the leaves in plastic bags after severing the petiole (personal communication). The exact technique used to measure Ψl in the Escalona et al’s (2002) paper was not given but based upon the above discussion and the results from the measurements of Ψl from their study, it can be assumed that they did not bag the leaf blade prior to cutting the petiole, probably resulting in erroneous values. In addition, since Ψl and Ψstem are highly correlated with one another in grapevine (Salón et al. 2005; Stevens et al. 1995; Williams and Araujo 2002; Williams and Trout 2005), it is not surprising that Patakas et al. (2005) did not find Ψl to be correlated with the SFTi/SFTl ratio since they did not measure Ψl as outlined by Turner and Long (1980). Alternatively, measuring grapevine transpiration via the sap flow technique could have underestimated grapevine water use resulting in unreliable values (Shackel et al. 1992; Tarara and Ferguson 2001).
It has been suggested that Ψl is not a sensitive water stress indicator for woody horticultural crops (Higgs and Jones 1990; Jones 1990; 2004; Noar 1998) or that Ψstem is a better indicator of plant stress than Ψl for deciduous tree crops (McCutchan and Shackel 1992; Shackel et al. 1997) or grapevines (Chone et al. 2001; Patakas et al. 2005). Others feel that ΨPD is a better indicator of grapevine water status (Cifre et al. 2002; Medrano et al. 2003; Schultz 2003). Jones (2004) pointed out that temporal fluctuations, such as passing clouds, make the interpretation of Ψl measured during the day as an indicator of irrigation need unsatisfactory. We assume that the same would be true for the use of Ψstem. Jones (2004) also mentioned that such an indicator requires a reference or threshold value to be useful in irrigation scheduling. In the arid southern San Joaquin Valley of California, cloud cover is the exception during the summer months. During the course of this study (from August 16 until October 6), there was only one day in which clouds were present. Thus, the measurement of midday Ψl and gs occurred at non-limiting light levels. A fully irrigated baseline of Ψl (and Ψstem) as a function VPD at the time of measurement has been developed for grapevines (Williams and Baeza 2007). This is similar to that developed for trees using Ψstem as a function of VPD (McCutchan and Shackel 1992; Shackel et al. 1997). It was shown here that the ETLYS/ETo ratio was more highly correlated with the difference in the calculated Ψl of a fully irrigated grapevine as a function of VPD and the actual value of Ψl measured on the vines in the lysimeter during the course of this study than with measurements of midday Ψl.
Flexas and Medrano (2002) suggested that stomatal closure is one of the earliest responses of plants to drought and the dominant limitation to photosynthesis at mild to moderate drought. Jones (2004) also indicated that over short time-scales, stomata are sensitive indicators of water deficits. The linear decrease in gs (and ETc/ETo ratio) with Ψl reported in this study would indicate that measurements of midday Ψl reflect the reductions in both individual leaf gs and whole plant transpiration due to soil water deficits. The same could be said for midday Ψstem in the study by Mata et al. (1999). Shackel (2007) also found that there was a linear relationship (R2 = 0.88) between gs and Ψl of Pinot noir grapevines. These results differ from conclusions by Lovisolo et al. (2010) who after a review of the literature stated that there is no clear relationship between gs and Ψl in grapevines and that gs is better correlated with ΨPD. The correlations between Ψl and gs in this study and those found by Shackel (2007) are better than those between ΨPD and gs reported by Lovisolo et al. (2010), Medrano et al. (2003) and de Souza et al. (2005).
Stomatal control of leaf transpiration can affect plant tissue water status in response to various environmental and soil factors (Jones 1998). Plant species may respond differently to these stresses and have been categorized as ‘isohydric’ in which stomatal control maintains midday Ψl similarly across varying soil water deficits and ‘anisohydric’ in which less stomatal control allows for midday Ψl to decrease in response to soil water deficits (Tardieu and Simonneau 1998). It is often assumed that grapevines (Vitis spp.) are isohydric (Medrano et al. 2003; Patakas et al. 2005) or near-isohydric (Cifre et al. 2005) due to the fact that midday Ψl doesn’t differ between vines that are ‘well-watered’ or ‘irrigated’ and those experiencing soil water deficits. If this were the case, then midday Ψl measurements could not be used in an irrigation management program. Others have found that some cultivars of V. vinifera should be categorized as isohydric while other cultivars anisohydric (Schultz 2003; Vandeleur et al. 2009). However, Soar et al. (2006) stated that grapevines may be categorized as generally anisohydric. Williams and Baeza (2007) concluded that Cabernet Sauvignon, Merlot and Thompson Seedless grapevines should be classified as anisohydric based upon significant differences in midday Ψl of well-watered vines compared to vines receiving no applied water or were being deficit irrigated (leaf blades were covered with a plastic bag prior to severing the petiole and putting both in the pressure chamber in that study when Ψl was measured). The data presented in this study would confirm that Thompson Seedless was an anisohydric cultivar. Midday Ψl progressively decreased subsequent to the irrigation pump being turned off, and it was highly correlated with decreases in both single leaf gs and the ETLYS/ETo ratio.
The diurnal patterns of water use for the lysimeter vines when being irrigated (Figs. 3 and 4) are typical to those one would expect for latent heat flux of a well-watered crop (Allen et al. 1998). These patterns coincide with diurnal solar radiation values similar to that shown in Williams et al. (2003b). Conversely, maximum hourly water use of the vines plateau earlier in the day during the first dry-down period and hourly water use began to decrease by 1200 h (PDT) at the end of the second dry-down period in this study. These diurnal patterns for both fully irrigated and stressed vines in this study are similar to those reported by Mata et al. (1999) on peach trees grown in a weighing lysimeter prior to and during a dry-down period, respectively. The diurnal patterns of water use during both the irrigated and non-irrigated periods of this study differ from diurnal patterns obtained in several studies where sap flow sensors were used to quantify grapevine transpiration (Lopes et al. 2004) even those for supposed, well-watered vines (Lu et al. 2003; Patakas et al. 2005; Yunusa et al. 2000). In many of those studies, sap flow is maximized between 0800 and 1000 h, remains constant thereafter, and then decreases rapidly later in the afternoon. It is also interesting to point out that grapevine transpiration in those studies differs only slightly from one day to the next even though evaporative demand changes significantly from one day to the next. The differences between the diurnal patterns of grapevine water use found in this study and Williams et al. (2003b) and those reported elsewhere using sap flow sensors would indicate that the latter technique may not accurately measure whole vine transpiration (Tarara and Ferguson 2001; Shackel et al. 1992) and casts doubt on values of grapevine water use others have reported.
The differences in diurnal water use of vines irrigated at ETc and that after the irrigation pump was turned off are similar to those reported by Williams and Ayars (2005a) for vines prior to and after the trunks were girdled (removal of the phloem from around the trunk), respectively. Girdling grapevines will reduce leaf gs (Harrell and Williams 1987; Hofacker 1978; Kriedemann and Lenz 1972; Roper and Williams 1989), and it will remain lower until the girdle heals (Williams et al. 2000). Therefore, diurnal and daily water use of girdled vines will be lower than vines that are not girdled (Williams and Ayars 2005a). It is thought that girdling reduces gs due to an accumulation of abscisic acid (ABA) in the leaves of grapevines (During 1978; Loveys and Kriedemann 1974; Williams et al. 2000). ABA has also been implicated in reducing gs of grapevines in response to soil and atmospheric water deficits and the classification of grapevine cultivars as being either isohydric or anisohydric (Schultz 2003; Soar et al. 2004, 2006). Interestingly, Ψl is greater for girdled grapevines compared to those that are not girdled due to a reduction in transpiration (Roper and Williams 1989; Williams and Ayars 2005a; Williams et al. 2000). Thus, a reduction in gs due to girdling will keep Ψl values higher than those of non-girdled vines, while a reduction in gs due to soil water deficits in this study did not increase Ψl. There is evidence that plant hydraulic conductance may also play a role in the response of plants to soil water deficits and their classification as isohydric or anisohydric (Franks et al. 2007; Lovisolo et al. 2010; Schultz 2003). This may account for the results presented in this study.
Schultz and Stoll (2010) consider a ΨPD value in the vicinity of −0.2 MPa as no stress for field-grown grapevines. Ojeda et al. (2001) have categorized the level of water deficit experienced by grapevines with a ΨPD value of −0.2 MPa or greater as ‘nil’, while the level of water deficit experienced by vines with ΨPD values between −0.2 and −0.4 MPa as ‘weak’. Studies that have examined water relations in grapevines often report ΨPD values for their ‘fully irrigated’ vines as being close to −0.2 MPa or lower (Chaves et al. 2007; Correia et al. 1995; de Souza et al. 2005). Data presented in this study indicated as ΨPD decreased from −0.07 to −0.19 during the first dry-down period, the ETLYS/ETo ratio decreased 40%. The ETLYS/ETo ratio from September 15 to October 3, second dry-down period, decreased from 0.75 to 0.39, a 48% reduction when ΨPD decreased to −0.29 MPa. A reduction in the ETLYS/ETo ratio of 40% at a ΨPD −0.19 MPa is considerable, not nil, while a reduction in the ratio of 48% at a ΨPD of −0.29 certainly isn’t a weak level of water deficit. In addition, grapevines with gs values decreasing from 200 – 500 mmol H2O m−2 s−1 to 150 mmol H2O m−2 s−1 are assumed to undergo a mild water stress with gs values between 50 and 150 mmol H2O m−2 s−1 considered a moderate stress (Lovisolo et al. 2010). In this particular study, a reduction in gs from ~500 mmol H2O m−2 s−1 to just less than 200 mmol H2O m−2 s−1 decreased the ETc/ETo ratio greater than 40%. It would appear that the current values of gs assumed by many to indicate no or minimal water stress in grapevine needs to be reexamined.
The majority of grapevines planted on a worldwide basis are located in semi-arid to arid regions, locations with little or no rainfall during late spring and summer and with minimal cloud-cover. The results presented in this study would indicate that the measurement of midday Ψl is a sensitive indicator of vine water status and a reflection of daily water use of grapevines under the arid conditions of the San Joaquin Valley. Measurements of Ψl have been successfully used to indicate the water status/irrigation needs of apple trees grown in an arid region (Peretz et al. 1984) or to schedule deficit irrigation in a semi-arid region of Spain (Girona et al. 2006). Grimes and Williams (1990) reported that yield was highly correlated with seasonal mean values of Ψl,gs and the crop water stress index (CWSI). They also found that Ψl and CWSI were highly correlated with one another (unpublished data). Williams et al. (2010a, b) found that both vegetative and reproductive growth of Thompson Seedless grapevines grown in the San Joaquin Valley were linearly related to seasonal, mean values of midday Ψl. Therefore measurements of midday Ψl (or Ψstem), or their departure from a fully irrigated baseline Ψl as a function of VPD, could be used in a vineyard irrigation management program or to validate new stress monitoring techniques under arid or semi-arid growing conditions.
This study was funded in part by grants from the American Vineyard Foundation, Viticulture Consortium and the California Competitive Grants for Viticulture and Enology. Dr. Baeza was supported by a grant from the Programa Nacional para la Movilidad de Profesores de Universidad del Ministerio de Educación y Ciencia de España.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.