*pertinent crop reporting districts for KS and MO
crop progress charts use most recent data
Data visualization charts are for exploratory analysis. Note the begining and ending dates for activity of interest for respective states (choosing dates beyond these will return errors in remaining tabs). Some states do not report fieldwork days suitable after specific dates; setting end dates past these dates will not allow additional fieldwork days in those cases. Rather than considering the number of days suitable for fieldwork each week, it is useful to consider how many days were available between specific dates in the past. The number of days suitable at a range of probabilities (20th, median, and 80th percentiles) are displayed in the chart on the left. The grey shaded area indicates the range of observed days suitable (between the maximum and minimum) recorded as historic extremes during specific weeks.
Per USDA, a 'suitable' day is one where weather and field conditions allowed producers to work in fields a major portion of that day. Dates are week ending on Sunday.
The green and blue lines on both charts represent the 15th percentile and average number of fieldwork days, respectively. The green shaded area on the histogram represents the 15th to 45th percentile fieldwork days. The map below is in testing beta phase and displays changes in fieldwork trends as number of days per decade for the date range selected above.
More detail available: Griffin and Barnes 2017. Available Time to Plant and Harvest Cotton across the Cotton Belt. The Journal of Cotton Science 21:8–17.
Graphs are for exploratory data visualization, users choose four states (be sure to select four different states)
NOTE: Harvest LP results only applicable for Arkansas until yield adjustment sets (YAS) for remaining states are available
The shadow prices is the answer to the question 'How much would I be willing to pay for an additional unit of resource?'. The shadow price associated with a specific resource indicates how much returns to fixed costs would increase by increasing the amount of that specific resource by one unit. Units can be 1-hour of machinery use or 1-acre of acreage. Note that the shadow value is pertient for the next one unit of resource, not more than one as the value may change.
NOTE: Harvest LP results only applicable for Arkansas until yield adjustment sets (YAS) for remaining states are available
NOTE: Yield adjustment set and penalties after hurricane
NOTE: Yield adjustment set and penalties after hurricane for multiple-pass harvesters
NOTE: Yield adjustment set and penalties after hurricane for single-pass harvesters
NOTE: Harvest MIP results only applicable for Arkansas until for remaining states are available
Test. 1) user enter planting dates, then model truncates YAS 2) Terry build 2-pass harvest system 2) Glen sending expected yields for 1st and 2nd picking 3) Glen sending modified YAS for 1st and 2nd picking 4) Glen sending modified YAS for single pass harvest system 5) how many bolls can I pick per second, based on 3 bales per acre, how many acres can I pick? field efficiency? graph small boll buggy
Multi-stage or multi-period yield potential by plantXharvest date for Arkansas: development in progress
Yield adjustment sets aka yield potential matrices are being developed for each cotton-producing state; potentially customizable by the user. Example data below are being updated by Urban et al. 2023
A matrix generator is a computer program that allows the user to input the linear-programming data in a convenient form by prescribing only a minimum amount of information. Normally, it takes advantage of the repetitive structure that characterizes most large linear-programming models, where many of the coefficients are +1, —1, or zero. The matrix generator fills in those coefficients automatically, leaving to the user the task of providing only the input of those coefficients the program cannot anticipate or compute routinely. Multistage or multiperiod models have the tendency to repeat a given submatrix of coefficients several times. The matrix generator automatically duplicates this submatrix as many times as necessary to produce the final model structure. Matrix generators also provide consistency checks on input data to help detect possible sources of error. page 186 of http://web.mit.edu/15.053/www/AppliedMathematicalProgramming.pdf
Economic parameters on this tab are not connected but are being implemented via linear programming results on the harvester tabs
Remove this tab before Version 1 release: currently 29 April 204 testing V-20240429 Space-Heathen
contact Terry Griffin twgriffin@ksu.edu for comments, questions, suggestions
previous releases
Space-Heathen April 2024
two-pass heathen July 2023
Hot Autobots July 2023
Autobot Candy October 31 2022
Changes and updates listed below since previous version
hurricane SP YAS modified for allowing harvest to begin occuring normally when hurricane arrives later in season 27-29 April 2024
space wx capabilities updated 26-27 April 2024
space wx tab added, days suitable decreased for week of radio blackout 26 April 2024
Kathy Draeger discussed space wx tab and radio blackout 26 April 2024
hurricane has YAS for SP now in addtion to MP YAS 26 April 2024
Terry and Glen decided to integrate Glen's 2-pass system into multi-pass tab rather than dedicated tab 03 October 2023
Terry added multipass graph for production, harvest, carry 02 October 2023
Beth and Terry identified problem with capacity, should be == and not <- directions on constraints, corrected 29 September 2023
Terry temporarily updated MIP capacity constraints by 1/((ac/hr)*(lb/ac)) in Amat and hsfw in bvec for RHS 26 September 2023
Beth and Terry temporarily removed MIP capacity constraints for testing September 2023
Beth and Terry pressure test MIP September 2023
Glen and Terry truncated carryover penalty at 25% instead of 100% on 2P and MP September 2023
Caleb and Terry tweaks on September 2023
Caleb and Terry corrected the pound per week constraint error in bvec on 30 August 2023
Caleb and Terry corrected the acre-hours per pound error in Amat on 28 August 2023
number of passes now has limit using MILP
tab for hurricane scenario, emergency swarm deployment, modifies YAS by date of hurricane
tab for Glen Rain's 2-pass harvest system created, analysis and output forthcoming
implemented suggestions made by Caleb Lindhorst, TAMU, August 2022 to present
implemented suggestions made by Ty Griffin, K-State, 2023 to present
corrections made to selective multiple-pass robotic harvest using MIP
added tab for selective multiple-pass robotic harvest using mixed integer programming model, 25 October 2022
improved LP of single-pass harvest systems
implemented suggestions made by Luke Fuhrer, University of Georgia, September 2022
Linear programming (LP) model for single harvest pass system is operational, August 2022
Corrected error regarding multiple harvesters DSFW per Glen Raines, July/August 2022
Explantory notes added in tabs: May 2020
Changed harvester width units from feet to inches. Should we leave harvester1 as feet or keep both in inches? per discussions and suggestions with Glen Rains during February 2020
Bookmark function added to save parameter settings and set-up. per discussions and suggestions with Glen Rains during February 2020
Change minimum speed from 0.5 mph to 0.1 mph
allow user to modify their own start and end dates per Wes Porter
plotly graphs can be downloaded as *.png; per discussions and suggestions with Glen Rains: March 2020
moved from free/limited site to permanent Shiny server owned by KSU so no annual fees: October/November 2020
Initial alpha version developed during 2019 and first beta version presented to precision cotton workshop at 2020 Beltwide
Changes and updates forthcoming although unsure if feasible/possible; send suggestions to Terry Griffin
Update USDA NASS data or consider using live API; operational as of 1 September 2020 but NASS server unreliable for production
Autonomous harvest can begin at 10%? open bolls, yield penalty matrix should begin then for each state
customize by users' location; i.e. users in Georgia experience state set = to Georgia. Try session$clientData https://shiny.rstudio.com/articles/client-data.html
Connect API to USDA NASS to auto-populate data? or just update data once a year? Kansas (up to 2017) and Missouri (up to 2016) no longer report DSFW by CRD but could update the crop progress stats for those states however CRD not available via API
Output CSV of graphs per Glen Rains
robot harvest speed are variable. When fewer bolls, harvester moves faster and slower when bolls more dense. Bolls picked per minute may be more consistent if ground speed variable per Glen Rains February 2020
table shows prominent acreages, like for 1200, 1300 etc. per Wes Porter
allow user to input their own DSFW numbers per Wes Porter; use a multiplier
yield and quality by harvest date economics
include something like picker efficiency or boll losses, selecting things like: % bolls open for harvesting (would be around 70% for current harvest start date, ), % bolls left after harvest, % bolls fallen from weather or equipment. per discussions and suggestions with Glen Rains during February 2020
provide date of event and percentage for destruction? For example, given other parameters (planting date, state, etc,), how much is lost on Sept 1 if 50% of open bolls are gone and 25% of plants were destroyed (cannot get anymore bolls opening) from hurricane, hail, tornado or just heavy rain. Atlantic Hurricane season is June 1 to November 30. Could be calculated separately from the data you already provide. Want to play around with this as it could be a major benefit to be able to harvest as bolls are opening and not be exposed for 2-3 months for weather events to destroy part of the crop. We could start with this and add a probabilistic component to come up with predictions. per discussions and suggestions with Glen Rains during February 2020
Project team, contributors, commenters, and reviewers
Terry Griffin, Project Lead, Kansas State University
Elizabeth Yeager, Kansas State University: LP and MIP
Ty Griffin, Kansas State University: testing and development
Jon Devine, Ed Barnes, Gaylon Morgan, Cotton Inc., various suggestions since inception
Glen Rains, University of Georgia, several comments February 2020 to present
Wes Porter, University of Georgia, commented 23 October 2019
Luke Fuhrer, University of Georgia, commented 2022
Caleb Lindhorst, Texas A&M University, commented August 2022 to present
Brian Mills, Mississippi State University, commented 27 August 2020
members of ASABE committee updating D497 commented
Gregg Ibendahl, Kansas State University
.
..
Instructions for use of tool
Contact: Terry Griffin Kansas State University, with any questions or comments