A Collaborative Phenology Modeling System to Enhance Crop Management on Vegetable Farms

Final Report for SW12-037

SW12-037 (project overview)
Project Type: Research and Education
Funds awarded in 2012: $203,610.00
Projected End Date: 12/31/2015
Region: Western
State: Oregon
Principal Investigator:
Nick Andrews
Email
Oregon State University
Expand All

Project Information

Summary:

Our goal with this project is to develop an online crop management tool that alllows vegetable growers to time farming and marketing activities using degree-days. Initial conversations with grower collaborators indicated that vegetable variety models and weed models would be useful. Some were also interested in the role of thermal time in nitrogen mineralization. In this project our emphasis was on vegetable models and usability of the Croptime Calculator.

We used a collaborative process to develop an online vegetable crop management interface (the Croptime website) that is easy to use. The Croptime Calculator has been formatted to work on desktop computers, tablets and smart phones. Model output can also be copied into Excel or Word files for farm records.

From 2013-2015 we collected enough crop development data to develop at least 60 vegetable models from 15 different crops and three weed models. By March, 2016 we had completed 11 new vegetable models (broccoli and cucumber) and three new weed models. By March 2017 we expect to have at least 50 vegetable models online.

In 2015 and 2016 we have conducted extensive outreach to students, farmers and other agricultural professionals. Interest in the project and interest in using Croptime models appears to be very strong among farmers, seed companies, other agricultural professionals and students.

Project Objectives:
  1. Participatory research and system development.

Before this grant, the OSU phenology modeling system was primarily a pest management decision tool. We worked closely with fresh and processed vegetable growers and seed company staff using organic and conventional methods to identify top priority weeds, vegetables and varieties to model.

These collaborators helped us identify website features that would make Croptime more useful as a decision-tool for vegetable growers. They also provided input as we developed the Croptime Growth State Guide (adapted from BBCH). Interested farmers, students and agricultural professionals also volunteered to beta-test the Croptime website.

  1. Online degree day planning tool.

Coop incorporated high priority features into the Croptime Calculator. He is also working with Andrews, Heinrich and Peachey to post new vegetable and weed models online and to document modeling results. With supplemental funding he has developed 7-month degree-day forecasts derived from NOAA climate models.

  1. Phenology model development.

Crops prioritized by Croptime collaborators include snap beans, broccoli, cabbage, carrot, cauliflower, cucumber, kale, lettuce, parsnip, sweet pepper, spinach, summer squash, sweet corn, tomato and winter squash. We have collected five or more data sets (crop development data at one site and planting date) for 68 vegetable varieties. This is sufficient to develop degree-day models for 50-68 variety specific degree day models.

We have introduced degree-day models that predict time to germinable seed production for three weed species: hairy nightshade (Solanum physalifolium), lambsquarter (Chenopodium album) and redroot pigweed (Amaranthus retroflexus). We are also publishing information about the influence of seasonal temperatures on nitrogen cycles on organic farms. Weed species, crops and varieties were prioritized by growers using a participatory process.

An OSU Extension Publication that discusses nitrogen mineralization and the influence of temperature is in preparation: Plant-available nitrogen supplied by mineralization: Vegetable cropping systems (western Oregon). Sullivan (author) has identified a model that predicts nitrogen mineralization from total soil N measurements, estimated decomposition rates and mean soil temperatures.

 

Introduction:

Calendar days to maturity do not reliably predict maturity dates between regions or planting dates. Time is easy to measure with calendars and easy to report in seed catalogs. Although degree-days are more accurate for predicting maturity, they are inherently more difficult to count and predict than calendar days. Our goal is to develop a user-friendly interface that links vegetable growers to vegetable and weed degree-day models so that they can adopt degree-day models as a more accurate decision tool. Degree-day models have been used since the 1960s; we are extending their use to vegetable growers.

We have developed an educational Croptime web-page on the OSU Small Farms website. The educational web-page also links to the Croptime Calculator which houses the degree-day models and is hosted by the OSU Integrated Plant Protection Center’s Pest and Crop Models website.

The Small Farms educational web-page has instructions for using the Croptime Calculator and fact sheets about using degree-day models. The site also includes information for potential collaborators interested in collecting data for new Croptime models. The Croptime Growth Stage Guide describes how to collect crop development data and provides a data collection template. It includes growth stage descriptions for the 15 vegetables we are currently working with. We are also working on an introductory fact sheet that describes the modeling methods we are using in this project.

We used a participatory process to identify priority vegetable crops and varieties, and weed species. Growers also helped us identify and describe critical growth stages of each vegetable crop for the Croptime Growth Stage Guide. Collaborating growers and seed companies also hosted about 200 Croptime vegetable and weed development observations.

The Croptime Calculator currently has eleven new vegetable degree-day models and three new weed models that predict first viable seed set. During the funded period of the project we collected enough crop development data (data-sets) for at least 50 vegetable models. Our modeling work is ongoing, but we expect to have 50 vegetable degree-day models for 15 different vegetable crops online by spring 2017. We also introduced new functions to the Croptime Calculator and have conducted usability tests which are helping to make the calculator more useful and intuitive for vegetable growers.

Approximately 985 farmers, students and agricultural professionals have been introduced to Croptime at conferences, workshops, field days, and in undergraduate classes. More than 140 people have received some hands-on training with the Croptime website. Andrews presented Croptime at the American Society of Horticultural Science annual meeting in New Orleans, LA on Aug 5, 2015. He has submitted a proposal to give an oral presentation at the National Small Farms Conference on Sept 20-22, 2016 in Virginia Beach, VA.

Unexpectedly we have also begun to develop undergraduate curriculum to teach students about degree-day modeling, and invite them to provide data sets for future Croptime vegetable models. Croptime educational slides are available online for interested instructors. We are also beginning to work with seed companies to include degree-day modeling information in their seed catalogs.

An OSU Extension Publication is also in preparation. Sullivan, Plant-available nitrogen supplied by mineralization: Vegetable cropping systems (western Oregon). This publication primarily addresses the effects of substrate (kind of organic matter) and temperature on the timing of plant-available N supply for crops. It will also provide case-study data illustrating timing of N supplied by N mineralization in our vegetable cropping systems.

Two peer-refereed journal articles are also in preparation.

Cooperators

Click linked name(s) to expand/collapse or show everyone's info
  • Nick Andrews
  • Jim Bronec
  • David Brown
  • Dr. Len Coop
  • Bob Egger
  • Jim Gill
  • Peter Kenagy
  • Scott Latham
  • Frank Morton
  • Dr. Jim Myers
  • Dr. Ed Peachey
  • Dr. Dan Sullivan

Research

Materials and methods:
  1. Participatory research and system development:

Through focus group meetings and one-on-one discussions with grower and seed company collaborators, we were able to identify priority vegetable varieties and weed species for the Croptime modelling research.

Noordijk and Andrews developed the Croptime Growth Stage Guide that is posted on the Croptime website. This guide is used by collaborators when collecting data for Croptime models. It describes how to collect crop development data for vegetable varieties and includes a field data sheet for recording field observations. We adapted the BBCH-scale to make it more user-friendly for collaborators and more relevant to vegetable growers in the Western region. Growers, collaborating scientists and student workers provided input into the Growth Stage Guide and reviewed draft versions.

In January, 2015 we conducted four usability tests using the preliminary Croptime website. Donna Bosworth-Andrews (previously with Microsoft, Inc.) trained OSU faculty and staff to conduct software usability tests and helped us interpret findings. Participants worked through a series of tasks. They were given little to no prompting. Andrews, Coop and Noordijk facilitated the tests and made notes when participants had trouble using certain parts of the website. Donna also helped us develop a “$100 survey” to prioritize features for incorporation into Croptime. This work was part of a supplemental Croptime grant funded by the Clackamas Extension Innovation program. Our goal with the Croptime website has been to make the system useful and intuitive for vegetable growers and other users.

The Croptime quick guide illustrates steps needed to use the Croptime Calculator, and the educational slides (link provided in the introduction) provide background information about degree-day models and Croptime. We are also developing a screen recording that will demonstrate how to use Croptime.

  1. Online degree day planning tool:

A new interface and back-end calculation system were developed to meet DD calculation needs of Croptime users. This allowed incorporation of several features that could not be programmed into the older DD calculation system, which is still in use for dozens of non-Croptime models. Guided in part by feedback from usability tests and user surveys, and partly by needs outlined in the original proposal, the new interface provided several important features needed for scheduling vegetable plantings. These include:

  1. Google-map selection of weather station,
  2. Pull-down menus for CROPTIME models
  3. Multiple start dates
  4. Spreadsheet-ready output (via cut and paste)
  5. Condensed output (showing data only for dates with actual events)
  6. Links to model development docs and model results preview
  7. Hand-held device compatibility
  8. Daylength calculations and critical threshold settings and markings (for daylength-sensitive species),
  9. Multiple options for an extended forecast (following the 7 day forecast). These include:
    1. NOAA based seasonal (7-month) forecast (work funded by USDA NIFA CPPM ARDP program), this was selected as the default option.
    2. Use last year’s data as an extended forecast (e.g. thus far the current year is similar to last year’s weather),
    3. Use data from the year before last as an extended forecast,
    4. Use a 10-year average, and
    5. Use a 30-year average.
  10. DOS code to help expedite the modelling process.

The interface has been well received during several hands-on workshops and feedback for additional desired features was gathered.

  1. Phenology Model Development:

We collected data from collaborating farmer’s fields in 2013 and 2014 with help from student workers. In 2014 and 2015 we collaborated with vegetable seed companies and Jim Myers (OSU vegetable breeder) to collect crop development data from their variety trials. We also collected data from our own variety trials at NWREC. Variety trials were the most efficient way to collect crop development data. In 2016 we are starting to collect crop development data with undergraduate students participating in organic student farms at Evergreen State College (Olympia, WA), Clackamas Community College (Oregon City, OR) and Oregon State University (Corvallis, OR). Professors teaching these classes believe that collecting Croptime data will help students learn observational skills.

Coop trained Andrews, Heinrich, Noordijk and Stoven in degree-day modelling methods. We drew heavily on the methods described by the University of California Davis: http://www.ipm.ucdavis.edu/WEATHER/ddconcepts.html.

Vegetable variety models:

At least eight data-sets (full season plant development observations) were collected to verify threshold temperatures for each vegetable crop. Thermal time to maturity on subsequent varieties of the same vegetable crop is being modeled using at least four data-sets. From 2013-2015 we collected 504 crop development data sets from 104 different popular vegetable varieties to support model development. We collected at least four data sets for 68 different varieties. Gerdts (student worker) developed a comprehensive list of weather stations that provided high quality weather data for all of the Croptime observation sites.

To date we have posted four transplanted broccoli degree-day models, one direct-seeded cucumber and six transplanted cucumber models. We will post seven sweet pepper models and six sweet corn models soon, and the rest of the models will be online by spring 2017. We continue to work on the modeling and documentation. We are using lowest co-efficient of variance to determine thresholds and thermal time to maturity. We are testing various degree-day calculation methods when verifying thresholds. So far all models except sweet corn are using single sine with a horizontal cutoff for upper threshold temperatures.

Weed models:

Aaron Heinrich and Ed Peachey developed DD models for hairy nightshade, lambs-quarter and redroot pigweed. These models predict first germinable seed set starting from emergence of dicotyledons. Three to five plants of each species were identified at the cotyledon stage on each farm/field. At approximately the 2-4 true leaf stage, 3-5 of the most vigorous plants were chosen for evaluation and competing plants around the model plant were removed. Key growth stages were monitored such as first flower for nightshade, or inflorescence emergence for pigweed and lambsquarters. Berries and pods were collected when seed began to mature. Collected seed was dried and stored at room temperature in coin envelopes for 4-6 months before germination. Seeds were treated with gibberellic acid or stratified to maximize germination. Following seed conditioning, the seeds were placed in petri dishes and sandwiched between paper discs moistened with deionized water. Only seeds with strong elongated radicles were counted as germinated. Threshold temperatures and thermal time to germinable seed production were modeled using the same techniques as the vegetable models.

Nitrogen: 

Funding limits prevented us from doing original work on nitrogen management. Nevertheless, Sullivan agreed to interpret his existing data and review the scientific literature related to the role of temperature in nitrogen mineralization.

This information may not result in an online nitrogen management tool, but it will inform growers about rates of nitrogen cycling during the growing season, which is year-round in western Oregon and Washington.

Research results and discussion:

Participatory research and system development:

During the first season (2013) we identified more than sixty priority vegetable varieties and six priority weed species. Growers agreed to collect crop development data on their farms, but it quickly became clear that despite their willingness, their work load didn’t allow time for them to monitor and record crop development. We also lacked clear growth stage descriptions for most of the vegetables we were working with. Since we had to draft and revise these growth-stage descriptions on an ongoing basis during the project it was much more efficient to collect all of the data internally. We did not anticipate this, so we secured supplemental local grants through the Clackamas Extension Innovation grant program to fund this additional labor cost. Andrews and Noordijk collected much of the data, we also hired student workers and recruited interested Extension faculty (i.e. Amy Garrett and Chip Bubl from OSU, and Jeremy Cowan from WSU) to collect additional crop development data. In 2013 and 2014 we collected data on host farms and at variety trials managed by OSU and Bejo Seed. In 2015 we collected data from variety trials in Aurora, Corvallis and Eugene, Oregon, because it was much more time efficient. In April 2016 we published the first edition of the Croptime Growth Stage Guide online (39 pgs). We are now using this Guide to collaborate with university horticulture classes and are hoping to continue collecting data this way even when grant-funding is low.

Usability tests and surveys from hands-on Croptime workshops provided valuable feedback on how to improve the Croptime Calculator. The usability tests, surveys and informal discussions informed Croptime Calculator development in 2014 and 2015. The current version of the calculator was introduced at conferences and workshops in the winter of 2015/16.

Supportive comments in evaluation surveys included: “Helpful in determining harvest and weeding schedule based on planting dates and crop type, super cool!”; “Can’t wait until it is fully populated. Excellent site.”; “It’s great; this will be an invaluable resource when it’s complete”. 75% of respondents at the Small Farms Conference and 84% of respondents at the NWREC workshop said that the current site is easy or very easy to use. One respondent said that Croptime “Needs a better interface” but unfortunately didn’t offer specific ideas for improvement. Several good suggestions were made to improve ease of use, such as “I didn’t recognize the orange headers as links. Maybe a “click here” feature” (we have addressed this issue); “more start dates” (we increased from one to four); “Running the model based on harvest date would be useful” (this is a priority feature for future work); “show days of the week”; “display average first and last frost dates” (these features have not yet been incorporated).

We used a “$100 survey” where participants “purchased” new website features with 100 pretend dollars. The following investments in new features were highly prioritized by 34 respondents in the hands-on session at the OSU Small Farms Conference (Feb, 2016) and NWREC (Mar, 2016) workshops:

Features

Total investment

Average investment

Ability to enter “harvest date” and have calculator predict best “planting date”

$945

$29

New crop or pest models not already included

$770

$32

Ability to set up a personal CROPTIME account (i.e. with saved settings for weather stations, varieties and start dates)

$440

$15

Incorporate day-length triggers for day-length sensitive species (i.e. spinach or weeds)

$340

$15

Display average first and last frost dates in model output

$290

$13

Irrigation scheduling (only asked at the NWREC workshop)

$125

$16

 

Online degree day planning tool:

Coop implemented several improvements to the Croptime Calculator based on feedback from growers, seed-company and University collaborators. These new features are listed under Materials and Methods. Ongoing feedback is identifying new functions that can be implemented with future funding.

Phenology model development

Coop trained Andrews, Heinrich, Noordijk and Stoven in degree-day modelling methods, checked the quality of our work, wrote DOS code to speed up the modelling process, and posted new models online as soon as they were available. Using data collected during this project, we expect to exceed our goal from the proposal of about 40 new vegetable models by spring 2017.

We published 11 vegetable varieties (broccoli and cucumber) and three weed models (hairy nightshade, lambsquarter and redroot pigweed) on the Croptime website in Feb, 2016. We expect to publish seven sweet pepper and six sweet corn models by June, 2016. At this time we can’t accurately predict the number of models that we have sufficient high-quality data. But by spring 2017 we expect to publish a total of at least 49-53 vegetable models on the Croptime website.

Model input summaries show model parameters (thresholds and thermal time to modeled growth stages). When more models have been posted online we plan to post a summary of model parameters on the website to help farmers and agricultural professionals in other states use the Croptime models on their local University agricultural weather and phenology modelling websites. The crops included in the project are listed in the Objectives section.

We published three weed models in early 2016. As with the vegetable crops, model parameters are posted on the Croptime website model output. %CV from cotyledon to first germinable seed was 11.4% for hairy nightshade, 9.4% for lambsquarter, and 9.3% for redroot pigweed.

The Extension Publication on nitrogen will introduce concepts, and illustrate them with case studies. Sullivan is also introducing this equation published by Gilmour (2009) that predicts N mineralization from soil organic matter decomposition.

N min = Soil N x [1-exp((-k)(TF))], where:

Nmin = PAN produced from soil organic matter (lb/acre/day)

Soil N = soil N (lb/acre, 0-12 inches)

K = daily OM decomposition rate, 0.0002 per day at 77 oF

TF = temperature factor based on Q10, equal to 1.0 at 77 oF

Research conclusions:

During the grant-funded portion of this project we have focused on developing the Croptime website, data collection, degree-day modeling, and outreach to farmers, students and peers. It is too soon to estimate adoption or the economic impact of this project except in very preliminary ways.

We plan to develop a survey that can be completed online when people click on the Croptime Calculator. This will allow us to measure future impacts of the Croptime project. We have used a similar approach to estimate the economic impact of the OSU Organic Fertilizer and Cover Crop Calculator: http://smallfarms.oregonstate.edu/calculator. The registration survey for the cover crop calculator feeds into a database at OSU which we use to estimate the economic impact of that decision-tool. The online Croptime survey will be written in Qualtrics and designed to estimate adoption rates and economic impacts.

In annual crops, weeds are most competitive when the crop is being established. Later in the season, weeds tend not to reduce yield or quality. However, if a significant weed in the field sets viable seeds, it increases the weed seed bank. Croptime weed models predict development of first germinable seed and help growers determine whether late season rogueing or cultivation is worthwhile.

Participation Summary

Research Outcomes

No research outcomes

Education and Outreach

Participation Summary:

Education and outreach methods and analyses:

The largest outreach tool and publication from this project is the Croptime website. So far we have published 14 new degree-day models for vegetable varieties and weed species. We expect to publish an additional 40 vegetable variety models by spring 2017. The website also hosts the Croptime Growth Stage Guide (Noordijk and Andrews, et al. 2016, 39 pgs), a Quick Guide with step-by-step instructions describing how to use the Croptime website, and 53 educational powerpoint slides have been posted online.

Approximately 985 farmers, students and agricultural professionals have been introduced to Croptime at conferences, workshops, field days and in undergraduate classes. More than 140 people have received some hands-on training with the Croptime website. Presenters have included Andrews, Noordijk, Coop, Heinrich and Sullivan.

  1. Croptime: online vegetable scheduling (scheduled for May 12, 2016) at the OSU Mid-Columbia Agricultural Research & Extension Center, Hood River, OR. This is for small-scale farmers and will include a presentation and hands-on opportunities to use Croptime.
  2. Croptime: online vegetable scheduling (April 26 and scheduled for May 17, 2016) at the Oregon State University campus. Two presentations with hands-on practice using Croptime for HORT 260: Organic Farming and Gardening. These undergraduate students are interested in collecting data for new Croptime models in 2016. 65 participants.
  3. Croptime: online vegetable scheduling (April 14, 2016) at Clackamas Community College in Oregon City, OR. A presentation with hands-on practice using Croptime for HORT-141: Urban Agriculture Practicum. These undergraduate students are interested in collecting data for new Croptime models in 2016. 15 students.
  4. Croptime: online vegetable scheduling (April 5, 2016) at Evergreen State College in Olympia, WA. A presentation with hands-on practice using Croptime for students enrolled in the Practice of Organic Farming program. These undergraduate students are interested in collecting data for new Croptime models in 2016. 25 participants.
  5. Introducing Croptime: online vegetable scheduling (Mar 14, 2016) at the OSU North Willamette Research & Extension Center. Intensive workshop with in-depth presentations and a hands-on session for participants to use Croptime. 26 participants.
  6. Introduction to Croptime (Mar 9, 2016) at the Winter Vegetable Field Day at the OSU North Willamette Research & Extension Center. 74 participants.
  7. Environmental influence on crop yield and quality by Ted Radovich, (Feb 19, 2016) in undergraduate Vegetable Crop Production Class (TPSS 401). 17 students.
  8. Introducing Croptime: online vegetable scheduling. Double session with presentations and a hands-on session (Feb 20, 2016) at the OSU Small Farms Conference. 73 participants.
  9. Croptime demonstrations (Feb 4-6, 2016) at the Organic Seed Alliance Conference in Corvallis, OR. We purchased an exhibitor table to demonstrate Croptime. More than 500 people attended the conference, we estimate that we demonstrated Croptime with about 40 people.
  10. Croptime: online vegetable scheduling (Jan 30, 2016) at the Horticulture Growers’ Short Course: Lower Mainland Horticulture Improvement Association and Pacific Agriculture Show. Abbotsford, BC. 40 participants.
  11. Croptime: online vegetable scheduling (Jan 12, 2016) at the North Willamette Horticulture Society annual meeting, Canby, OR. 163 participants.
  12. Croptime: scheduling vegetable crops with degree-days (Nov 14, 2015) at Tilth Producers of Washington 2015 Annual Conference. 120 participants.
  13. Croptime degree-day models for vegetables (Sep 17, 2015) at the Vegetable Variety Field day at OSU North Willamette Research & Extension Center. 45 participants.
  14. Croptime degree-day models for winter squash (Sep 15, 2015) at OSU Veg Farm. Part of the Northern Organic Vegetable Improvement Collaborative (Jim Myers PI) field day. 20 participants.
  15. NWREC Learning Farm, Croptime and Vegetable Variety Trials (Aug 26, 2015) at the OSU North Willamette Research & Extension Center. Field tour for the International Trials Conference. 45 participants.
  16. Decision Tools for Vegetable Growers: Nutrient Management and Crop Scheduling (April 18, 2015) in Corvallis. Presentation at International Association of Students in Agricultural and Related Sciences USA National Summit. 40 participants.
  17. CropTime: Vegetable Degree-days (Jan 13) in Canby. Presentation at North Willamette Horticulture Society Organic Crops Day. 139 participants.
  18. Croptime: Vegetable Degree-Days (Oct 21, 2014) in Corvallis, OR. Presentation at the Horticulture Departmental Seminars. 18 participants.
  19. Croptime introduction (May 2, 2014) at NWREC. Presentation for the Oregon Academy of Nutrition and Dietetics. 20 participants.

Andrews presented initial findings and early models for broccoli, cucumber, sweet corn and sweet pepper at the ASHS annual meeting in New Orleans on Aug 5, 2015: https://ashs.confex.com/ashs/2015/webprogramarchives/Paper20969.html.

Dave Muelheisen at Evergreen State College in Olympia, WA asked Andrews to introduce Croptime to his class. Dave thought the crop development data required to create degree-day models for vegetable varieties provides an excellent opportunity to teach his undergraduate students good observational skills on vegetable crops. He has incorporated Croptime into his curriculum, and students in his program are planning to collect degree-day data for future Croptime models. Andrews also taught similar classes at Clackamas Community College (Koniescka) and Oregon State University (Cassidy). We hope that this educational benefit will gradually enhance adoption and economic impact, and that students will collect data that is useful for future model development.

Sullivan is publishing an OSU Extension publication that discusses nitrogen mineralization from soil amendments and soil organic management with reference to thermal time.

We are drafting two journal articles. One describes Extension methods and will be submitted to the Journal of Extension, the other describes the crop and weed modeling work and will be submitted to HortTechnology in 2016 or 2017. Our need to collect three years of crop development data before modeling efforts could begin has made it impossible for us to publish journal articles earlier in the project.

Education and Outreach Outcomes

Recommendations for education and outreach:

Areas needing additional study

Collaborating researchers, growers and workshop participants have suggested several areas for future work:

  1. Increase the number of vegetables and varieties being modeled.
  2. Extend Croptime to winter vegetables grown in the open and in high tunnels.
  3. Develop software to expedite the modelling process.
  4. Add new weed models.
  5. Develop models for cover crops.
  6. Incorporate WSU’s irrigation scheduling tool into Croptime: http://weather.wsu.edu/is/.
  7. Add new user-interface features to the Croptime website, for example:
    1. Ability to set harvest date and have the model solve for planting/transplant dates.
    2. An alternative condensed output whereby events (rather than dates) are in column 1 and the predicted dates of events are in remaining columns for each planting date.
    3. Calendar format data output (such as with Google Calendar), etc.
  8. Enhance outreach to growers and other agricultural professionals.
  9. Publish Croptime and degree-day information in commercial seed catalogs.
  10. Extend Croptime models to agricultural weather and modelling websites at other interested Universities.
Return to Project Overview
Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.