We all know that climate and weather play a critical role in the economic success or failure of commercial fruit production. However, weather changes in recent years have caused growers and researchers to question what the future climate might hold for fruit production. In the Great Lakes region, winters have been milder, as evidenced by a change in frequency of ice formation on Grand Traverse Bay (see chart "Years"). We have also observed snow melt and fruit tree growth earlier in spring. As a result of these warming trends, the average date of tart cherry bloom in Traverse City, Michigan, now occurs seven to eight days earlier in spring than it did 30 years ago.

In 2002, the unimaginable happened: Michigan experienced record warmth in mid-April, which advanced tart cherry bud development enough that a comparatively typical subfreezing wind the following week caused substantial freeze damage. This combination of events virtually eliminated the tart cherry crop in the state. Production in northwest Michigan fell from an average of 145 million pounds to just 1 million pounds. Michigan produced its smallest crop of tart cherries since USDA began keeping records in 1925.

Pileus Project

Evidence of warmer winters and earlier spring bud development, paired with the freeze event, prompted researchers at Michigan State University to undertake a study entitled the "Pileus Project." This project was designed to assess the potential impact of climate change on tart cherry production in the Great Lakes region. The primary objectives of the Pileus Project first required researchers to create quantitative models that simulate Montmorency tart cherry phenological development and predict the relationship between climate variability and crop yield. The model strategy was to hold all aspects of tart cherry production constant except weather/climate in order to investigate potential impacts of weather.

Researchers first looked to historical information from 1960 to 2003 to help develop the simulation models. Historical production and weather data were analyzed with the other major components of the models that include phenology, cold injury, leaf area development, water balance and use, disease risk, and yield.

Based on the analysis of historical information, three significant weather-related factors were found to be associated with yield. The dominant factor affecting tart cherry yield is freeze injury to developing flower buds during spring.

The second factor influencing crop production in a given year is the effect of weather on pollination. The study found that yield is reduced as the number of wet days increases during bloom.

The third factor associated with yield is total precipitation during later stages of the previous growing season, i.e., more late-season rain improves the probability of a good crop the following year. This third factor was more weakly associated with yield than is freeze injury or weather during bloom. This relationship between late-season rain and crop yield was a rather unexpected finding in the Great Lakes region, and this result may influence future regional irrigation management strategies.

Due to the significant weather effect during the early growing season, it was necessary to develop a phenological model that would equate Montmorency flower bud development with growing degree days. The model simulated the critical period from the beginning of the growing season through petal fall.

A mathematical model was then developed that incorporated the three significant weather factors with the bud development phenology model to predict yield based on weather events.

Climate models

Pileus Project researchers used this information to project potential future impacts of climate change on tart cherry production based on four different global atmospheric models that simulate climate change. To help determine the relative uncertainty associated with the projections, a range of differing techniques and methodologies were used, along with the four separate climate models, to create a large number of potential future climate scenarios. Over 60 scenarios including daily maximum and minimum temperatures and precipitation were developed at each of 15 selected locations across the Great Lakes region.

With the large number of scenarios, the results provide a range of predictions for each study parameter. All scenarios suggest that growing degree-day accumulations will increase throughout the twenty-first century. In the cool fruit-growing area near Traverse City, growing degree-day base accumulations (41°F) are predicted to increase from 100 to 700 (average 350) growing degree days above mean historical levels by the middle of the century. By late century, growing degree days are predicted to increase from 300 to 1500, with the average increase at approximately 830 growing degree days.

The length of the growing season across the region (defined as the period between the last freeze of the spring season and the first freeze of the fall) is also expected to increase. By midcentury at East Jordan (near Traverse City), projections for change in growing season range from a decrease of 10 days to an increase of 30 days, with the mean of the scenarios suggesting an increase of 15.5 days. By the end of the century, the scenario range is an increase of 10 to 65 days, with an average of 35.5 (or 5 weeks) more frost-free days.

Bud development is also projected to follow the trend toward earlier development throughout the century. Projections about changes in precipitation and water balance are less definitive, so less can be concluded from these simulations.

Researchers combined the tart cherry yield prediction model with the climate change scenarios. The result is that more of the scenarios suggest increases in the fraction of viable buds (i.e., less cold injury) than decreases (see chart "Projected Change") for both mid and late century. However, the results do not suggest a strong positive difference. Regardless, given recent past events, the results were not what some of us were expecting. It increases our hope that the unusual conditions of 2002 will remain a fluke of nature rather than a preview of events to come.

For more information, including access to a suite of interactive decision-making tools, visit the project Web site at: www.pileus.msu.edu.