Back to top

Preharvest Prediction of Yeast Assimilable Nitrogen in Finger Lakes Riesling using Linear and Multivariate Modeling


Research In Plain English provides brief, non-technical summaries of journal articles by Cornell faculty, students, and staff.
(2013) American Journal of Enology and Viticulture

Authors: Mark A. Nisbet, Timothy E. Martinson and Anna Katharine Mansfield
Summary by Amanda Garris


Nitrogen is an essential metabolite for yeast and is often the limiting factor for fermentations. Without sufficient yeast assimilable nitrogen (YAN), fermentations can stall, but too much nitrogen can lead to microbial instability, spoilage and production of carcinogenic ethyl carbamate. Measuring YAN at harvest is not always practical because of time and equipment constraints, so our goal was to develop a convenient and accurate  preharvest measure which would predict YAN at harvest,  providing winemakers with a tool to make appropriate nitrogen additions at the start of fermentation.


Grapes were sampled from 60 Riesling vineyards in the Finger Lakes. Samples were taken from 24 vines per site, and the same vines were sampled in 2010, 2011 and 2012. The sites chosen captured a range of soil types and levels of vine vigor. Samples of 200 berries per site were collected at veraison, two weeks before harvest and at harvest. Berry samples were weighed and crushed, and pH, soluble solids and titratable acidity were recorded. YAN was calculated by measuring the two forms of nitrogen utilized by yeast: ammonia and primary amino nitrogen. In addition, standard soil fertility and soil heath measures were analyzed. Three methods of statistical analysis were tested for their ability to predict YAN at harvest from the veraison and preharvest samples.

What we learned

  • In the vineyards studied, juice samples were generally deficient in yeast assimilable nitrogen, with an average concentration of 92.0 mg/L.
  • The data suggests that 95% of vineyards would have a YAN concentration between 29 mg/L and 190 mg/L, and therefore winemakers adding only 120 mg/L supplemental nitrogen would ensure that less than 5% of samples would have a concentration below 140 mg/L YAN and only 0.1% would have a concentration above 400 mg/L YAN.
  • Using 120 mg/L supplemental nitrogen, the average concentration in wine made from Finger Lakes Riesling would be 206 mg/L, very close to the recommended dose. Supplementation at this level would reduce the cost of nitrogen supplementation and lower the risk of excess nitrogen. The simplest method to accurately estimate harvest YAN uses preharvest YAN or ammonia measurements and a straightforward equation (see box.)  However, as with any prediction, there will be some uncertainty for new observations. Therefore we have included a prediction interval of +/- 48 mg/L and +/- 43 mg/L for preharvest measurements of ammonia and YAN, respectively.
  • More complex models provided slightly more accurate predictions, but they require more analysis to obtain the values and would likely be less useful to winemakers.​
YAN equation

The bottom line

Measuring ammonia two weeks before harvest requires a single test which can be performed for low cost and with relative ease, and it provides flexibility and predictive power for winemakers who want to estimate YAN at harvest.

Amanda Garris is a freelance writer based in Geneva, NY.