GRAPEGROWING

The Limitations of the Winkler Index By Patrick L. Shabram

he Winkler Index or Winkler Scale is Methodology of the index days in the month for each month from April a standard for describing regional Modern-day calculations of GDD most often to October, then summed for the entire grow- Oct for in the United utilize daily accumulations of degree days. That ing season [∑Apr monthly((T_mean-50)•30)]. TStates. Developed by A.J. Winkler is to say that they sum the total degrees of aver- Further, a 1998 assessment of temperatures and M.A. Amerine at the University of Cali- age daily temperatures above 50° F (10° C) for in the city of Sonoma, Calif., determined that Oct 31 fornia, Davis in the first half of the 20th every day from April 1 to Oct. 31 [∑Apr 1 daily Amerine and Winkler’s original calculations century, the index was constructed to corre- (T_mean-50,0)]. If the average temperature for may have been simplified to account for only late quality with , focusing on a given day is 75° F, then the total degree days 30 days in each month.7 Hence, GDD originally viticulture. Wine-producing regions added to the total sum for that specific date calculated in 1944 may have had four fewer of California were broken into five climatic would be 25° F. Any average below 50° F con- days figured into the equation than what mod- regions using heat summations above 50° F, stitutes zero GDD for the given day. ern assessments typically apply. or growing degree days (GDD). Heat summa- An initial problem with some calculations The discrepancy between daily accumula- tions are a way of looking at accumulated is how “average” is determined. In most cases, tion of degree days and monthly accumulation temperatures over a given time period. De- average is calculated as the mean of the lowest of degree days is most pronounced in early- spite the common usage of the Winkler Index, recorded temperature and the highest re- season and late-season numbers, when mean the classifications offer greater uncertainty corded temperature, so a high temperature of daily temperatures may be below 50° F. As a than the system suggests. 85° F and a low temperature of 65° F would simple example, assume that the average tem- An obvious drawback to the Winkler Index produce 75° F [(Tmax-Tmin)/2]. Another method perature for each day from April 1 to April 15 is the focus on temperature alone. Winkler’s to calculate average temperatures is to take is 48° F and mean temperature for each day groundbreaking resource “General Viticulture” the mean of all temperature readings for a April 16 through April 30 is 54° F. Using the notes the influence of “rainfall, fog, humidity given day. If a weather station reports hourly daily accumulation method, each day from and duration of sunshine,” but Amerine and temperature readings, the mean of those tem- April 1 to April 15 would have a degree-day Winkler’s work and subsequent research found perature readings would constitute the average total of 0, while each day from April 16 to April that temperature plays the greatest role in the of the 24 daily temperature readings, while if 30 would have a degree-day total of 4, for an development of wine . As the focus of a station reports in 15-minute intervals, the April total of 60 degree days. With an average Winkler and Amerine’s wine- research average temperature would be the mean of the monthly temperature of 51° F, using the was on California viticulture, their climatic 96 daily temperature readings. The simple monthly average would yield a GDD total of regions had the luxury of ignoring precipita- difference in methodology for calculating aver- 30 degree days for the month. tion, as little rain falls during the growing age temperatures can have an impact on GDD Even with consistency in methodology, season. Indeed, Winkler et al. in “General Vi- readings. A recent study of five weather sta- other factors impact the overall GDD total, and ticulture” suggest that the vinifera grape “is not tions in the AVA, for example, hence the climatic region assigned to that total. suited to humid summers, owing to its suscep- found variations in 10-year average GDD as Equipment and placement of weather stations tibility to certain fungus diseases and insect high as 206 based on the methodology uti- are considered, although most government pests that flourish under humid conditions.”9 lized.6 As many weather stations have software and research weather stations, and even most Advances in viticulture have expanded the generating GDD figures for the user, the meth- range of successful commercial vinifera pro- odology deployed is not always obvious. Fur- FIGURE 1: GDD (°F) AT CAMINO, duction, while more recent research notes the ther, many software programs will vary on CALIF. USING TWO METHODOLOGIES importance of wind on photosynthesis rates.2 GDD calculations, including but not limited to Winkler et. al. also point out the importance growing season, threshold (e.g., 50° F or some Daily Monthly of ongoing refinement of local variations, stat- other number), and how temperatures below Accumulation Accumulation ing, “It is hoped that refinements will be de- that threshold are handled. Some weather Lowest (1998) 2889 2752 veloped so as to delimit subregions within the station software programs, for example, will Highest (2017) 4034 3932 present regions, thereby ensuring the greatest not include temperatures below 50° F in the 30-Year Average 3439 3300 potential for quality when the most favorable average temperature calculation. 10-Year Average 3583 3465 climatic subregion for a given variety is planted Daily accumulation methodologies, how- 5-Year Average 3821 3639 to that variety.” ever, are not consistent with the methodology 30-year average GDD based on 2017-1987 exclud- Therefore, the Winkler Index was designed deployed by Winkler and Amerine. Rather, ing 1988 because of incomplete data in 1988. at a state scale, and not necessarily intended Winkler and Amerine used monthly means. for smaller viticultural subregions or single Specifically, the mean monthly temperature 10-year average GDD based on 2017-2008. 5-year vineyards. above 50° F was multiplied by the number of average based on 2017-2013.

108 &VINES Collector’s Edition GRAPEGROWING commercial weather stations in- offers some historical context for of Camino. Diamond Springs is a growing districts become more stalled in vineyards, are assumed the Camino area in general. Thirty much newer station, installed defined at subregional levels and to be placed to limit variables that total years of data were assessed Sept. 20, 2010, so a 30-year aver- as we test new environs for com- may impact climatic data. The from 2017 to 1987, with 1988 age is not possible. The five-year mercial viticulture, local com- duration of the data (i.e., number removed from the analysis, as average at Diamond Springs parisons are valuable. of years of data), however, can data were incomplete. The results based on daily accumulations is have a significant impact on num- in Figure 1 show that daily accu- 3,997 degree days compared to Duration of high bers. Typically, meteorological mulations led to higher GDD to- 3,821 at Camino. The seven-year temperatures normals (i.e., what is considered tals than using monthly means. daily accumulation average Another consideration with Win- to be the average weather for a (2011-2017) at Diamond Springs kler climatic regions is the dura- particular location) are based on Variations in averages is 3,913 compared to 3,679 at tion of high and low temperatures. 30-year averages, updated every Both the daily accumulations Camino. Every year with the ex- GDD calculations based on mean new decade (e.g., 2010-1981, and monthly accumulations to- ception of 2017 shows higher daily temperatures treat high and 2000-1971, etc.). Finding com- tals at the Camino station are GDD at Diamond Springs than at low temperatures equally, even if plete data sets covering 30 years relatively consistent with the the Camino weather station the high temperature is reached in a given area can be difficult, so total found in Winkler et al., (2017 shows Camino 30 degree for only for a few minutes, but the GDD averages used in viticulture although the numbers are 139 days higher than Diamond low temperature persists for sev- are most commonly given in some degree days different, based on Springs). Even if you could not eral hours. Areas impacted by significantly shorter duration. methodology. If a grower had give a good long-term average afternoon marine inversions may Further complicating data sets been given a single year’s data, GDD for Diamond Springs, it be especially susceptible to briefly is what happens around a given however, and that year hap- would be safe to say that the cli- maintained high temperatures. weather station during that 30- pened to be either the year with mate at the Diamond Springs Figure 3, created for a study done year period. Urban development the lowest GDD (1998), or the station is warmer than at the in the Livermore Valley AVA, can create warmer conditions year with the highest GDD Camino station. As viticultural shows the cumulative number of that, combined with a changing (2017), very different conclu- climate, lead to migration from sions could have been reached FIGURE 2: GDD TREND APRIL 1-OCT. 31 AT CIMIS 13 one climatic classification to an- about the climate of Camino. other over the course of historical The averages for the five most CAMINO 1984-2017 data sets. Further, academic stud- recent years show both a warmer 4500 ies are often done in Celsius, with climate than the 30-year average the Winkler climatic regions con- would suggest and greater vari- verted to Celsius equivalents. For ance between methodologies. 4000 example, Region II is converted to Keeping with standards of clima- 1,389-1,667° C (from 2,501- tology and meteorology would 3500 3,000° F), giving the impression discredit the five-year average as that the climate-region break- insufficient, as weather varies sea- downs are based on something son to season, and climate is a 3000 more accurate than the conve- long-term averaging of weather. nience of even 500 degree day Yet getting five years of data is intervals. (Since the original re- often much more achievable than 2500 1980 1985 1990 1995 2000 2005 2010 2015 2020 gion breakdown is in Fahrenheit, obtaining 30 years of data. The Using daily accumulation method based on daily mean temperature. Data for 1986 and this article is a discussion of climate is also warming. Figure 1 and 1988 are incomplete and are not included in the graph. those regions, Fahrenheit is used demonstrates the long-term tem- throughout this article unless oth- perature trend in GDD for Camino, erwise noted.) with every academic indication FIGURE 3: HOURLY READINGS AT HIGH TEMPERATURE To demonstrate how varied that this trend will continue. The n 7/16/16 n 7/13/16 GDD numbers can be for a given prospects are very likely that on- n 7/15/16 n 7/12/16 n 7/14/16 n 7/11/16 location, data were assessed from going climatic conditions at 16 a California Irrigation Manage- Camino will more likely reflect n 7/10/16 14 ment Information System (CIMIS) GDD above the 30-year average. station at Camino in El Dorado Which Winkler Index climate re- 12 County, Calif. This station was gion is most appropriately applied 10 chosen because it is one of the to Camino? Any interested party 8 longest continuously operating could have a choice between Re- 6

CIMIS stations, having been in- gion III or Region IV, or if they Hourly Readings stalled Oct. 19, 1982. Data from were really looking to cherry-pick, 4 this station are also nearly com- Region II or Region V. 2 plete, especially after 1989. The Comparative analysis is where 0 station is not near a growing a more appropriate usage of GDD Pleasanton Point Santa Oakville Brentwood Manteca Modesto urban center. Camino is listed in may come into play. How does San Pedro Rosa “General Viticulture” (Region III Camino compare to other nearby Weather Stations with 3,400 GDD). This station is areas? A CIMIS station is also Adopted from Shabram, 2017. All data collected from California Irrigation Manage- not the same weather station used located at Diamond Springs, ap- ment Information System. Temperature reading within 0.3°C or 0.5°F of the daily in the Winkler calculations but proximately nine miles southwest maximum temperature were counted from July 10-16, 2016.

Collector’s Edition WINES&VINES 109 GRAPEGROWING hourly readings at or near the daily high tem- known to be relatively free of coastal fog. Variations in growing season peratures for several CIMIS weather stations Many areas transitioning from coastal to in- A final consideration is the April 1 to Oct. 31 in Northern California over a one-week period land may show GDD numbers similarly influ- growing season used in GDD calculations. In in July 2016. Cumulative high temperatures enced by brief durations of high temperatures. areas of the West Coast, bud break typically were maintained for a shorter period at Santa Comparatively, similar GDD summations for occurs in March, sometimes as early as Febru- Rosa and Pleasanton, two areas known for areas at higher latitudes or greater elevations ary, while harvest dates in August and Septem- variable summer marine intrusions, than at may show greater total hours near high tem- ber are more common than in October. Grape Point San Pedro, Brentwood and Modesto. peratures or longer periods of solar radiation. variety and microclimatic conditions in indi- Point San Pedro, at the immediate coast, has Using average air temperature rather than vidual vineyards also play a role. greater consistency in marine stratus layers daily mean temperature may create more ac- A review of 50% bud break from 2013 to than Santa Rosa and Pleasanton, which experi- curate comparative numbers, assuming that 2018 and harvest dates from 2013 to 2017 was ence fluctuations in coastal fog. Brentwood the methodology is consistent across the areas conducted in seven vineyards in the Livermore and Modesto, both inland from the coast, are being assessed. Valley AVA, three planted to , three planted to and one planted to . The average 50% bud break ranged from March 28 to April 5, depending on vineyard, for the three Cabernet Sauvignon vineyards, but the Chardonnay vineyards had average dates of March 14 to March 15. The Sauvignon Blanc vineyard ex- perienced 50% bud break around March 26. LIQUID Harvests for Cabernet Sauvignon approxi- mated Oct. 8 to Oct. 21, depending on the vineyard. Chardonnay saw average harvest dates of Sept. 20 to Oct. 7. The vineyard with NITROGEN Sauvignon Blanc saw average harvest dates around Aug. 28. In many cases, bud break began prior to the April 1 growing season start, DOSING while harvest was most often completed prior to the Oct. 31 growing season end. regardless of your closure preference The rationale for calculating heat summa- tions from April 1 through Oct. 31 are not clear, at least not in Amerine and Winkler’s best- known works on the subject. The 50° F thresh- old is based on shoot growth, so April 1 may represent an approximate estimation of bud break, but harvest for many varieties occurred, even in the first half of the 20th century, prior to Oct. 31. Amerine and Winkler also calcu- lated heat summations from bloom to harvest and noted the role of heat summations in the 30 days preceding harvest, therefore not limit- ing their assessment to just April through Oc- tober. Heat summations based on bloom and harvest are only possible when grape produc- tion already exists within an area and would be difficult to use as a tool to match varieties Minimize dissolved oxygen to an area yet to see vines. Differential timing Extend shelf life and lengths of growing seasons across North America further complicate the equation. Purge O2 from empty bottles

Purge O2 from headspace Is there a better index? Other indexes also exist for assessing tempera- Since 1958 ture. The Heliothermal Index, also known as the (HI), was introduced by Pierre Huglin in 1978. HI is the April 1 to Sept. 4 Barten Lane, Woburn, MA 01801 30 sum of the mean of the daily mean temper- T 781-933-3570 ate above 10° C and the high temperature F 781-932-9428 above 10° C, multiplied by a coefficient indica- [email protected] tive of the latitude. Other indices include Aver- age Growing Season Temperatures (GST) and vacuumbarrier.com Biologically Effective Degree Days (BEDD), along with several other formulas for calculat- ing heat summations.

110 WINES&VINES Collector’s Edition GRAPEGROWING

A 1999 study by Jorge Tonietto and Alain paring growing areas, but GDD is just one aspect makes a good comparative tool when consis- Carbonneau suggests that a multicriteria system of climatic analysis that alone gives an incom- tency in methodology and comparison are made of calculating HI, a dryness index (DI), and a plete picture (as are other indices designed to within regional locations. cool night index (CI) are needed to account for summarize heat summation). The methodology, potential water balance in soil and nocturnal daily climatic patterns, period assessed and Patrick Shabram is a geography professor at Front temperatures in grape development.8 Gregory length of growing season can all tell a different Range Community College in Colorado and an expert Jones and others used PRISM climate models picture of mesoclimates that might otherwise on American Viticultural Areas (AVAs). He conducts to assess spatial climatic distribution across appear to be similar by Winkler designation. geographic research on wine-grape growing regions California, Idaho, Oregon and Washington Ideally, each viticultural region would find a and is a consultant for growers interested in establishing based on the four most commonly used indices climatic classification system that best suits the new AVAs or modifying boundaries to existing AVAs. in viticulture.3,4 The results suggested that GST unique characteristics that define it, while re- and GDD had the greatest correlations, but HI serving simpler indices like the Winkler Index The references for this article are and BEDD demonstrated better differentiation for broader generalizations. Alternatively, GDD available online at winesandvines.com of climate types. A similar study by Rosalyn Francine MacCracken and Paul R. Houser used the PRISM model to predict how well these four indices, along with a modified version of the GST accounting for the common length of a growing season (Modified-GSTavg or Mod- GSTavg), performed in U.S. states east of Cali- All American fornia, Oregon and Idaho.5 The MacCracken and Houser study found all common indices to be Containers less than perfect at predicting the viability of successful viticulture when comparing the indi- ces’ results with existing viticulture. Mod- YOUR RIGID PACKAGING GSTavg showed more promise as a predictor of SOURCE successful viticulture by taking into account the length of the growing season. The Jones et al. study and the MacCracken and Houser study looked at viticulture on a super-regional scale, albeit using climatic Your wine. models with 400m to 400m resolution. One could argue that the Winkler Index was never meant to be utilized on a super-regional scale, Our bottles. but rather as a statewide guide specific to California. The original 1944 paper looked only at California locations, with Winkler later introducing other U.S. and international loca- What an tions into the scale for comparison. Winker et al. are clear that the regional scale was not meant to be definitive at a subregional scale: excellent “These divisions into climatic regions should be considered as general demarcations.”9 pairing. Winkler et. al. suggest more extensive work be conducted on more local areas and specifi- cally note the work of Robert Sisson, who, dur- ing his 35-year career as viticulture farm advisor to Sonoma County from the 1950s to the 1980s, created a climate classification system unique to aacwine.com Sonoma County. Sisson’s model broke the county into Marine, Coastal Cool and Coastal Warm climate types, based loosely on heat summations of hours spent between 70° F and 90° F. While much work has been done to delineate subre- gions of viticultural areas, both formally through the creation of new, smaller AVAs and informally through the identification of subregional dis- tricts by winegrowing associations, work to All American Containers, a Veritiv company create climatic classification systems at more local levels unfortunately is scarce. Use of the Regions I-V designations of the WINDSOR, CA • SAN LEANDRO, CA • SIMI VALLEY, CA Winkler Index presents several limitations and KALAMA, WA • PLANO, TX • BELLE VERNON, PA • TAMPA, FL MIAMI, FL • ATLANTA, GA • BRANCHBURG, NJ errors. GDD calculations, on which the Winkler Index is based, can be a valuable tool for com-

Collector’s Edition WINES&VINES 111