Scheduling Vegetable Planting for Winter Protected Production in the Midwest

Environmental Monitoring in High Tunnels and Low Tunnels on Three Farms

Environmental records for 2015-2016 and 2016-2017 are summarized in Figs. 2 through 10.

Crop Planting and Harvest Dates and Yield

Crop planting and harvest dates from 2015-2016 and 2016-2017 are summarized in Table 5. The growing degree days and daily light integral accumulated from planting to harvest covered a wide range even within crop types (data not shown). Yield for the 2015-2016 season is presented in Table 6.. As mentioned in the methods section, harvests were determined in part by market requirements and so yield reported for some crops does not represent their full potential. Generally, yield decreased with later planting dates in the same environment.

Mathematical relationships between daily solar radiation and average daily temperature outside, and average daily temperature inside an unheated high tunnel were identified and are illustrated in Figure 11. A comparison of the air temperature in unheated tunnels predicted by the relationship and the actual temperatures measured in Oct. - Dec. 2016 is shown in Fig. 12. The predicted and actual accumulation of growing degree days in these structures is illustrated in Fig. 13.

We constructed a graph to illustrate how these results could be used to create the basis for a scheduling tool, assuming a consistent relationship between growing degree days (GDD), daily light integral (DLI) and harvest maturity can be described (Fig. 14).

For three locations in Indiana we used historical records of typical daily solar radiation and normal average daily temperature  (obtained from online sources) to predict average daily temperature an unheated, double-poly high tunnel. For a hypothetical crop that would consistently take 600 GDD (base 40°F) from planting to harvest, the graph shows days to harvest for various planting dates, for the three locations. Similar graphs could be constructed for more locations and crops requiring different numbers of GDD and/or DLI accumulation. The formulas used to create the graphs could be used as the basis for a scheduling tool. Actual planting and harvest dates for spinach grown in high tunnels are also plotted on the graph for comparison. The actual data points follow a pattern similar to the points for the hypothetical crop, but they don't line up exactly, and some line up more closely with the southern Indiana location than for northern Indiana.

A significant amount of work is needed in order to develop a scheduling tool using this approach. Importantly, we need to better define the relationship between GDD, DLI and harvest maturity for different crops. We can do some exploration with the data we collected, but data from more controlled experiments covering a wider range of planting dates would be helpful. We also need to quantify, or estimate, the degree of uncertainty that surrounds information provided by the tool.

Spinach Row Cover Study

Yield from the row cover study is shown in Table 7. Results differed at the three locations, but led to the overall conclusion that in a relatively mild winter like the one when this study was conducted, row cover generally led to greater yield compared to no row cover, but it did not matter whether the row cover was supported on hoops or laid on the plants like a blanket. Farmers reported that the blanket system was easier and did not reduce spinach quality. Based on these results, farmers said they would use the blanket system for spinach.

Table 7. Spinach yield, unharvested weight, total biomass, and plant stand in three row cover treatments on three Indiana farms, 2016-2017.

¹LOC: UH=unheated high tunnel; HHT=heated high tunnel with thermostat set to 25-30°F; LT=low tunnel.

²TRT: NRC=no row cover; RCB=row cover blanket on crop; RCH=row cover supported by hoops above crop. Medium-weight row cover, 0.9 oz./sq. yd.

³For each farm, statistics were used to compare 1) the average of the two UH row cover treatments vs the UH NRC treatment, and 2) the UH RCB vs UH RCH. Different letters after the means indicate statistically significant differences. Uppercase and lowercase letters indicate significance at P<.05 and P<.10, respectively.

We did not get a complete set of air and soil temperature data for the spinach trial due to technical problems. The wireless tags did not always connect and so there are periods with no data. The manually downloaded loggers were sometimes not downloaded before they wrote over previously recorded data. The daily average soil temperature for CB and EP is presented in Fig. 15. Soil temperature under the two row cover treatments was similar and during the coldest periods about 5 degrees warmer than without row cover.

Radish Planting Date

Results of the radish planting date study are illustrated in Figs. 16-18.

Radish developmental rate slowed as planting was delayed. This made it easy to predict a harvest date based on a planting date, as shown in Figure 16 and in Table 8. More work with this data is needed to understand how much these days to harvest would change under warmer or cooler conditions, and whether using temperature and/or light information would lead to a better prediction of harvest date. Total yield per bed-foot trended downwards with later planting dates (Fig. 17), but factors other than planting date could have influenced this trend: the planting density varied; the precise bed length was not known; and the number of harvests varied from 1 to 3. Yield for plantings up to Oct. 17 ranged from 1 to 1.6 lb., or 1.2 to 1.9 bunches, per foot of bed length. The size of individual radishes is illustrated in Fig. 18, which shows the weight per radish (including leaves) vs. days from planting to harvest. Most of the radishes were harvested when they weighed .03 to .04 lb., or about 1/2 to 2/3 oz. Radishes in the Sept. plantings grew about 1/4 oz. in a week.

Table 8. Expected days to harvest and first harvest date for 'French Breakfast' radishes planted Sept. 1–Nov. 22 in an unheated high tunnel, based on data from Eden Prairie Farm, Indiana, 2016-2017.

Baby Leaf versus Multi-leaf Lettuce

Baby leaf lettuce was first harvested on Dec. 22 and multi-leaf on Dec. 26. Fig. 19 shows the lettuce plantings about a week after the first harvest. Yield in a 90-ft. bed was estimated at 68 lb. for the baby leaf and 140 lb. for the multi-leaf.

For winter production in high tunnels, key recommendations derived from this project include:

• Maximize sunlight for winter high tunnels, so it will be warmer inside
• Remove row covers on sunny days
• Be prepared to ventilate to reduce temperature on sunny days in winter to avoid excessively high temperatures
• If crop won't be damaged, consider laying row covers right over the crop instead of supporting the row cover on hoops to save labor and expense
• Plant crops that will be harvested multiple times (e.g. kale) as early as possible (August or early September) for best yield.
• Keep records of planting and harvest dates on your operation to fine-tune schedules for your farm and markets. Find a simple and reliable method to keep records, e.g. notes on wooden stakes stuck in ground for each crop
• Use transplants to reduce the time from planting in tunnel to harvest
• For reliable survival of less hardy crops (e.g. lettuce or Swiss chard) and greater and more consistent production of hardy crops, provide supplemental heat

Environmental Monitoring Equipment

A few comments about the different environmental sensor and logger systems may be helpful for others interested in environmental monitoring. The equipment from Onset (RX 3000 logger and associated sensors) was the most expensive and the most complex to set up. It took a significant amount of time to install and remove, largely due to the the large number of sensors and cables. Cables were strung through conduit or unused irrigation tube or hose to protect from chewing animals and damage from farm tools; this also added time to installation. Establishing the wifi connection also added to the set-up time. After installation the systems worked well and consistently, but not entirely without problems. To troubleshoot and repair required travel to the farms. The company provided readily accessible technical assistance by phone, and quickly sent replacement parts when the issue was a faulty part. It was helpful to be able to see the data online in a graphical format with only an hour delay; some partner farmers made use of this feature more than others.

The WatchDog temperature/relative humidity logger, and attached PAR or temperature sensors from Spectrum Technologies were easy to install and reliably recorded data. It is also possible to view current measurements on a small screen on the logger. If not downloaded often enough, data is written over and lost, and that proved to be a problem a couple of times. The length of time before a download is needed will depend on the number of sensors connected, how often data is logged, and the memory size of the logger. Downloading requires proprietary software from Specmeters, and downloaded files must be converted in order to be read by common spreadsheet software. The proprietary software offers options for specialized reports that provide calculations for growing degree days, DLI, and other useful computations based on the recorded data. The button loggers from this supplier similarly must be downloaded at intervals in order to view the data. They do not have any screen where current data can be seen.

The wireless tag sensors used at some locations in the spinach row cover trial are less expensive than the two systems mentioned above. At some locations we were operating close to the limit of the wireless system, and data was not consistently transferred. Also, the wireless signal can drift over time and the tag then loses connection with the base station and no longer transmits data until it is made to reconnect. The wireless tags are not designed to withstand water, and so can stop working if they are sprinkled. A smartphone app designed for the system displays data and can also receive alerts when the temperature is out of a range set by the user, and alert via voice or sound. Data viewable online or through the app can be emailed to or downloaded by the user.