Predicting Optimum Time of Alfalfa Harvest

Dairy July 19, 2010|Print



Optimum timing of alfalfa harvest is critical to obtain high-quality forages for lactating dairy cows. Three methods used to estimate neutral detergent fiber (NDF) of alfalfa were evaluated in a field research project conducted during 2000 on 35 farms in Michigan. The three methods evaluated are growing degree-days (GDD), predictive equation for alfalfa quality (PEAQ), and scissors-cut samples. Predicted NDF based on the three prediction methods were compared to NDF of chopped alfalfa both before and after ensiling in laboratory silos. Based on results from this project and previous research, PEAQ and GDD methods adequately predict NDF of first cutting alfalfa; however, only the PEAQ stick should be used for second cutting. The GDD is not reliable when there is inadequate soil moisture, which often occurs during second cutting growth. Neither GDD nor PEAQ adequately predicted NDF for third cutting alfalfa. The scissors-cut method can be considered for predicting NDF for the third cutting.

Please check this link first if you are interested in organic or specialty dairy production


Alfalfa is an important forage for dairy cows because it provides fiber that effectively stimulates chewing while also providing energy and protein for milk production. There is an optimum quality for alfalfa that should be fed to dairy cows. Quality can be too high or too low for maximum milk production.

The measure of fiber most commonly used to balance diets of lactating dairy cows is NDF. The optimum concentration of NDF for alfalfa is 40%. Alfalfa containing 40% NDF allows reasonable grain concentrations in the diet while maintaining adequate NDF concentrations. The protein concentration of alfalfa with 40% NDF is usually moderate (approximately 20% of DM), and additions of low-protein grains such as corn allow flexibility in diet formulation for ruminally undegraded protein while avoiding excessive protein concentrations (Allen, 1997).

Delaying alfalfa harvest increases NDF percentage and reduces protein concentration. More grain will be required to increase energy density and decrease the NDF concentration (and filling effect) of the diet. In addition, more supplemental protein will be required to meet the cows’ protein requirements, and DM intake and milk production will be reduced.

Methods of Predicting NDF

Several methods recently have been proposed to predict timing of first cutting alfalfa harvest based on NDF concentration:

  • growing degree-days (GDD, base 41°F),
  • predictive equation for alfalfa quality (PEAQ) based on plant height and stage of maturity, and
  • scissors-cut samples.

At the farm level, procedures need to be easy to use, require minimal time, and provide relatively accurate predictions of the NDF concentration of the alfalfa that is to be fed.


The GDD calculation for alfalfa is based on the minimum and maximum daily temperature beginning on March 1 and using a base of 41°F. The daily calculation is:

[(max. temp. + min. temp.)/2] - 41

The total GDD is the sum of the positive daily growing degree-day values across days beginning March 1.

Based on research at Michigan State University (MSU) (Allen and Beck, 1996), alfalfa averages 40% NDF at about 750 GDD. An NDF concentration of 35% is approximately equivalent to 600 GDD. It takes a total of about 970 GDD to reach 45% NDF.

These GDD predictions of NDF concentration are for pure alfalfa stands. Grass matures earlier than alfalfa. Consequently, significant amounts of grass in alfalfa fields will result in higher NDF concentrations if the forage is harvested based on GDD estimates of NDF.

The GDD also is most effective in predicting harvest of first cutting alfalfa. Predicting NDF concentration using GDD cannot be done when there is inadequate soil moisture because GDD accumulates with little or no response in plant growth. Consequently, GDD has been used only for first cutting alfalfa harvest.


The PEAQ method (Hintz and Albrecht, 1991) is based on an equation that uses the length of the tallest alfalfa stem and the stage of the most mature alfalfa plant (will likely be two different plants) in the area sampled. The current modified PEAQ method uses a scale of three stages of maturity (late vegetative, bud, and flower). Measuring sticks, calibrated for the three plant maturity stages, are used to obtain estimates of NDF.

While obtaining PEAQ estimates of NDF, producers can scout their alfalfa fields for winter injury, disease development, insect damage, and weed encroachment (Sulc et al., 1999).Good sampling technique is critical to obtaining reliable NDF estimates. It is important to obtain NDF estimates from the PEAQ method at five or more locations across the field.

As with GDD, the PEAQ method was developed for pure stands of alfalfa. The NDF estimates from PEAQ will not account for weeds or grasses in the stands. The PEAQ is not reliable for estimating NDF when alfalfa is very short (longest stem is less than 16 inches) or very tall (longest stem is more than 40 inches).

Scissors-cut Samples

Scissors-cut samples provide a direct measurement of NDF in the collected plant material. Sampling technique is critical. A representative sample must be obtained from across the field. Sample handling is also important in minimizing respiration losses prior to the sample arriving in the analytical lab. In addition, errors can occur with near infrared reflectance spectroscopy (NIRS) analysis of scissors-cut samples because equations for fresh alfalfa are not generally available (Sulc et al., 1999).

Description of Project

In 2000, we conducted a field research project in Michigan to compare different methods of predicting alfalfa NDF concentrations over first, second, and third cuttings. The methods compared included GDD (base 41°F), PEAQ, and scissors-cut samples.

The project consisted of samples and data collected at alfalfa fields in 35 locations throughout Michigan, including five locations in the Upper Peninsula. Daily maximum and minimum temperatures were collected with electronic data loggers at each field. Temperatures were recorded every 10 minutes beginning on March 1 and continued through the duration of the project (approximately mid-August).

We compared the NDF predicted from these samples with the NDF analyses of chopped alfalfa both before and after ensiling in laboratory silos. Immediately prior to cutting the alfalfa field, the PEAQ stick was used to predict NDF and the scissors-cut sample was taken. The field-wilted alfalfa was sampled immediately prior to chopping. This sample was manually “chopped” and either dried within 24 hours or immediately ensiled in laboratory silos. The scissors-cut, chopped, and ensiled samples were analyzed for NDF concentration by wet chemistry procedures at the MSU Department of Animal Science.


The NDF concentration of the ensiled alfalfa ranged from 35 to 46% for first and second cuttings. Third cutting NDF ranged from 35 to 52%.

The NDF concentration of ensiled alfalfa samples was predicted adequately by all three methods for the first and second cuttings. Although there was little difference between GDD and PEAQ for first and second cutting alfalfa, we do not recommend using GDD for predicting NDF concentration of second cutting alfalfa. There is often inadequate soil moisture for second cutting growth, and we believe the GDD method is not reliable in these conditions.

The scissors-cut method was the only method that adequately predicted NDF for third cutting. It should be noted that these samples were handled under controlled conditions. The cut samples were chilled immediately and delivered to the lab within 24 hours of collection.

The error associated with the methods was slightly lower for PEAQ compared to GDD and the scissors-cut method for the first and second cuttings. When PEAQ was used to predict NDF, about two-thirds of the samples were predicted within +2.3 units of NDF for first cutting and within +2.8 units of NDF for second cutting. When GDD was used, about two-thirds of the samples were predicted within +2.6 units of NDF for first cutting and within +3.1 units for second cutting. The corresponding measurements of error for the prediction of NDF from scissors-cut samples were 2.4 units of NDF for first cutting and 3.0 units of NDF for second cutting.

There was good agreement between NDF concentration of the fresh chopped and ensiled samples. The regression equation is:

NDF%-ensiled = 10.8 + 0.72 * NDF%-fresh chopped, with an R2 of 0.55, RMSE (root mean square error) of 2.2, and P <0.0001.

Note, however, that these samples were handled under ideal conditions. It is likely that there would be greater differences in NDF between fresh chopped and ensiled samples with standard ensiling procedures on the farm.

What We Recommend

The following recommendations are based on the results of this project and previous research:

  • Use the PEAQ stick or GDD to predict NDF for first cutting alfalfa. Only the PEAQ stick should be used for second cutting alfalfa. Neither PEAQ nor GDD is recommended for third cutting alfalfa.
  • Begin cutting alfalfa at 40% NDF (750 GDD, base 41°F) for upright silos and 38% NDF (680 GDD, base 41°F) for horizontal silos. Start even earlier for horizontal silos if it takes more than a week to finish harvesting.
  • The GDD and PEAQ methods cannot be used for fields containing grass.
  • Fields containing grass should be harvested first. Start with the fields with the most grass, and finish with the purest alfalfa fields.
  • Consider using the scissors-cut method for fields containing grass and for third cutting alfalfa. Shipping samples to the analytical lab by next-day delivery will help to minimize deterioration in sample quality. Wet chemistry analysis is most appropriate for scissors-cut samples.


We would like to express our appreciation to the 19 MSU Extension Agents and MSU Research Associates who diligently collected data and alfalfa samples. We also acknowledge the efforts of Dave Main who conducted the laboratory analyses. Lastly, we thank the 35 dairy producers across Michigan who cooperated in this project by allowing us to collect alfalfa samples and temperature data from their fields.

Author Information

Kathy Lee, MSU Extension, Michigan State University
Mike Allen, Department of Animal Science, Michigan State University
Rich Leep, Department of Crop and Soil Sciences, Michigan State University


Allen, M. 1997. Harvest alfalfa for optimum quality. Michigan Dairy Review. May issue, p. 11-13. Michigan State University, East Lansing.

Allen, M., and J. Beck. 1996. Relationship between spring harvest alfalfa quality and growing degree days. p. 16-25. In Proc. Twenty-sixth National Alfalfa Symposium, East Lansing, Mich., March 4-5, 1996. Certified Alfalfa Seed Council, Davis, Calif.

Hintz, R.W., and K.A. Albrecht. 1991. Prediction of alfalfa chemical composition from maturity and plant morphology. Crop Sci. 31:1561-1565.

Sulc, R.M., K.A. Albrecht, V.N. Owens, and J.H. Cherney. 1999. Update on predicting harvest time for alfalfa. p. 167-177. In Proc. Tri-State Dairy Nutrition Conf., Fort Wayne, Ind., April 20-21, 1999. The Ohio State University, Columbus.