This description relates to estimation of weather information using field-driven data obtained from multiple remote weather stations.
Farmers, producers, and agronomists use farm management systems to support their agronomic management and agricultural planning process. Farm management and agricultural management systems, commonly require a variety of data inputs to perform necessary calculations for the agricultural management cycle. Many of these inputs are categorical variables and properties that support agricultural management life cycles, including and not limited to, soil properties, elevation, seed type, crop variety, nutrient applications, weather, and so on.
Meanwhile, when each categorical variable is examined, there are many additional properties that comprise the categorical variable. With particular interest in the attributes that comprise weather properties, which may include, temperature, humidity, precipitation, wind speed or direction, to provide a few examples. Whenever available, the weather information collected directly from the agriculture field will drive more accuracy in the decision-making process.
One approach to collect weather data directly from the field is to install weather stations at all desired locations. However, by installing hardware at every desired location, this may impose a substantial cost. Therefore, a method to retrieve and leverage data from already existing weather stations would be beneficial. Further, a method that can approximate weather data for locations that do not include a weather station would be beneficial.
Described is a method for predicting a weather state by interpolating weather information from weather measurement systems located in or near a query location such as an agricultural field. The method is more efficient, less computationally expensive, easier to interpret and more accurate than more traditional weather interpolation methods.
In particular, a weather prediction model accesses a number of indicator weather states from known and nearby weather measurement systems. Each weather measurement system is at a measurement location and configured to determine a weather state at the measurement location. The weather prediction model predicts a weather state for a query location using the weather states determined by the weather measurement systems. Predicting the weather state at the query location may consider the number of weather measurement systems near the query location, their geometrical configuration, and the values that are measured by each such weather measurements system. The values measured by the weather measurement systems are used to predict a weather state at the query location. Using the wearer prediction model reduces the need to have a weather measurement system installed on every desired location in an area.
Predicted weather states at the query location may impact decisions that agronomists, producers, farmers, or farm managers make throughout the year. For example, this the weather prediction model may be applied to historical data that has already been collected, determine a weather forecast, and a farmer may choose a farming process based on the forecast.
The figures depict various embodiment for purposes of illustration only. One skilled in the art will readily recognize from the following discussion that alternative embodiments of the structures and methods illustrated herein may be employed without departing from the principles described herein.
I. Introduction
This method seeks to predict a weather state at a query location, or locations, in an area by interpolating results obtained from one or more weather measurements systems in the area. An area is some amount of geographic area that may include weather measurement systems and agricultural fields. Herein, a weather state is a quantification or measurement of some aspect of the weather such as, for example, precipitation, humidity, temperature, pressure, wind speed, etc. Thus, a predicted weather state is a predicted value of a weather state at a particular location in the area. Location and area centric weather information (e.g., weather states) is obtained from weather measurement systems or other sources of data to drive better and more efficient interpolated results for any missing locations within the area.
II. System Environment
A client system 110 is any system capable of executing a weather prediction model 112 to predict a weather state at a query location lq. The client system 110 may be a computing device, such as, for example, a personal computer. Network system 120 may also be a computing device, such as, for example, a set of servers that can operate with or as part of another system that implements network services for facilitating determining predicted weather states. Network system 120 and client system 110 comprise any number of hardware components and/or computational logic for providing the specified functionality. That is, the systems herein can be implemented in hardware, firmware, and/or software (e.g., a hardware server comprising computational logic), other embodiments can include additional functionality, can distribute functionality between systems, can attribute functionality to more or fewer systems, can be implemented as a standalone program or as part of a network of programs, and can be loaded into memory executable by processors.
In one example, a client system 110 is operated by a user responsible for managing crop production in an agricultural field within the area, but could be operated by any other user. The user of the client system 110 inputs a query location lq into the weather prediction model 112 and the weather prediction model 112 predicts a weather state for that query location lq in response. Generally, the query location lq is the location of the agricultural field, or a portion of the agricultural field, managed by the user, but could be any other query location lq. In some instances, the query location lq may be the location of the client system 110.
A client system 110 is connected to a network system 120 via a network 140. The network system 120 facilitates the weather prediction model 112 accurately predicting a weather state at the query location lq. In various examples, the network system 120 may access indicator weather states from weather measurement systems 130 in the area. The network system 120 can provide the indicator weather states to the client system 110 such that the weather prediction model 112 can predict a weather state at the query location lq. In some examples, the network system 120 (or the client system 110) may store any of the indicator weather states in a datastore. Stored indicator weather states may be accessed by weather prediction model 112 to predict a weather state at a query location lq. In some examples, the weather prediction model 112 is executed on a network system 120 and a client system 110 accesses the weather prediction model via the network 140.
A weather measurement system 130 is any system or device that can provide indicator weather states to the network system 120 and client system 110. In some instances, a weather measurement system 130 is a system or device capable of measuring and/or quantifying an aspect of the current weather (i.e., current weather states). For example, a weather measurement system 130 may be a weather station operated the National Weather Service, but could be any other weather measurement system 130. In other instances, a weather measurement system 130 may be an external system that stores previously measured weather states (i.e., historical weather states). For example, the weather measurement system 130 may be a database that stores historical records of weather in the area as indicator weather states. In either example, a weather measurement system may provide indicator weather states to network system 120 or client system 110. Notably, while system environment 100 illustrates two weather measurement systems 130, the system environment 100 can include any number of weather measurement systems 130.
The network 140 represents the communication pathways between systems in the environment 100. In one embodiment, the network is the Internet, but can also be any network, including but not limited to a LAN, a MAN, a WAN, a mobile, wired or wireless network, a cloud computing network, a private network, or a virtual private network, and any combination thereof. In addition, all or some of links can be encrypted using conventional encryption technologies such as the secure sockets layer (SSL), Secure HTTP and/or virtual private networks (VPNs). In another embodiment, the entities can use custom and/or dedicated data communications technologies instead of, or in addition to, the ones described above.
Within the system environment 100, each indicator weather state is associated with a measurement location lm. Each measurement location lm is a distance dm away from the query location lq. Generally, the measurement location lm of an indicator weather state is the location at which a weather measurement system 130 determined the indicator weather state. The distance dm between the query location lq and the measurement location lm can be determined by the weather prediction model 112 when predicting a weather state.
III. Predicting a Weather State
The client system 110 uses a weather prediction model 112 to predict a weather state for a query location lq. The weather prediction model 112 receives a query location lq as input and provides a predicted weather state at the query location lq as output. When predicting a weather state, the weather prediction model 112 may request and receive indicator weather states from network system 120 to facilitate predicting the weather state. Network system 120 may access the indicator weather states from a storage database of the network system 120 or from weather measurement systems 130 as previously described.
To begin, a weather prediction model 112 receives 210 a request to predict a weather state at a query location lq in an area. In this example, an operator of the client system 110 inputs the query into the weather prediction model 112 and initializes the request. Here, the query location lq is an agricultural field in an area and the operator is a person responsible for managing crop production of the agricultural field. The area also includes a number of weather measurement systems 130 that can provide indicator weather states for the weather prediction model 112. To demonstrate,
Returning to
Returning to
To demonstrate,
In some examples, the method 200 may determine a region 410 before receiving weather indicator states. In this case, only indicator weather states from weather measurement systems 130 within the region 410 are provided to the weather prediction model 112.
Returning to
Returning to
Weather prediction model 112 may determine the position of sectors within a region 410 based on the location of the proximal weather measurement system 510. In one example, a first sector of the determined sectors is positioned within the region such that a line from the proximal weather measurement system 510 to the query location lq 320 approximately bisects the first sector. The remaining sectors are equally spaced about the region 410 based on the location of the first sector. For example, if the region 410 is a circle, the weather prediction model 112 may partition the region 410 into six equally sized sectors. Each of the six sectors have an arc of approximately 60 degrees. In this example, the first sector is placed within the region 410 such that a line between the proximal weather measurement system 510 and the query location lq approximately bisects a sector such that 30 degrees of the sectors arc is on each side of the line. The remaining sectors are placed around the first sector to complete the circle.
To demonstrate,
Returning to
The representative location lr for a sector is separated from the query location lq by a representative distance dr. That is, dr=|lq−lr|. In one example, the representative location lr for each sectors 610 is located within the sector 610 such that a line connecting the representative location lr and the query location lq approximately bisects the sector 610. In other examples, the line may not bisect the sector.
In some cases, the weather determination model 112 can determine the representative distance dr based on the sector 610. In one example, in the sector 610 including a proximal weather measurement system 510, the weather prediction model 112 determines the representative distance dr for that sector 610 is the distance between the measurement location lm of the proximal weather measurement system 510 and the query location lq 320. Thus, the representative location lr is the measurement location lm of the proximal weather measurement system 510. In the other sectors, the weather prediction model 112 determines a representative distance dr that is a harmonic average of the distances between the query location and the measurement locations lq of the weather measurement systems 130 in that sector. Thus, in this case, the representative distance dr may not coincide with a measurement location lm of a weather measurement system 130. In other examples, the weather prediction model 112 can determine a representative distance dr using any other technique to average distances.
To demonstrate,
Returning to
The predicted weather state is any weather state that can be predicted by indicator weather states. For example, if all of the weather measurement systems 130 have an indicator weather state indicating a current amount of rain fall, the weather prediction model generates a representative weather state at a representative location lr 710 indicating rain fall for each sector 610. The weather prediction model 112 interpolates the representative weather states at the representative distances dr 712 of the representative locations lr 710 to predict an amount of rainfall at the query location lq 520. Of course, this is just an example of predicting a weather state. Weather prediction model 112 can use any indicator weather states described herein to predict a state.
Weather prediction model 112 outputs the predicted weather state to the operator of the client system 110. The operator of the client system 110 may use the predicted weather state to apply real-time decision making for the current agricultural life cycle in the agricultural field. The predicted weather state may be stored on a database of client system 110 or network system 120.
IV. Additional Model Outputs
In some examples, predicted weather states can be combined with other predicted weather states to generate a predicted weather map of the area. The predicted weather map can include any number of zones such that the predicted weather map represents a zone-by-zone map (or table) displaying current, historical, and/or predicted weather states in the area. The weather map includes no empty zones because of the predicted weather states.
Additionally, in some examples, current, historical, and/or predicted weather states may be combined about a sector 610 or zone in an area 310 to provide a holistic view of the sector 610 or zone. For example, if the weather states in the sector indicate current rain fall, high humidity, and a large amount of historical rain fall, the holistic view may include “flood risk.” In another example, if the weather states in a zone indicate no rain fall, high temperatures, a low amount of historical rain fall, the holistic view may include “drought.”
The predicted weather states may be analyzed to predict a future trend. For example, weather prediction model 112 can use predicted weather states from a previous growing season, or seasons, to predict a weather state for the current season.
In another example, a predicted weather state may be used by a machine in a field to take an action. For example, a predicted weather state indicating drought may be sent to a boom sprayer in a field and the boom sprayer may increase the amount of water provided to plants in response. Other examples of a farming machine utilizing a predicted weather state are also possible.
V. Additional Configuration Considerations
Likewise, as used herein, the terms “comprises,” “comprising,” “includes,” “including,” “has,” “having” or any other variation thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, article, or apparatus that comprises a list of elements is not necessarily limited to only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
In addition, use of the “a” or “an” are employed to describe elements and components of the embodiments herein. This is done merely for convenience and to give a general sense of the invention. This description should be read to include one or at least one and the singular also includes the plural unless it is obvious that it is meant otherwise.
Finally, as used herein any reference to “one embodiment” or “an embodiment” means that a particular element, feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. The appearances of the phrase “in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment.
Upon reading this disclosure, those of skill in the art will appreciate still additional alternative structural and functional designs as disclosed from the principles herein. Thus, while particular embodiments and applications have been illustrated and described, it is to be understood that the disclosed embodiments are not limited to the precise construction and components disclosed herein. Various modifications, changes and variations, which will be apparent to those skilled in the art, may be made in the arrangement, operation and details of the method and apparatus disclosed herein without departing from the spirit and scope defined in the appended claims.
This application claims the benefit of priority to U.S. Provisional Application No. 62/558,643, filed Sep. 14, 2017 which is incorporated herein by reference in its entirety for all purposes.
Number | Name | Date | Kind |
---|---|---|---|
6549852 | Hanson | Apr 2003 | B2 |
6651005 | O'Neall | Nov 2003 | B2 |
7702597 | Singh et al. | Apr 2010 | B2 |
8065124 | Xu et al. | Nov 2011 | B2 |
8412419 | Seamon et al. | Apr 2013 | B1 |
8917336 | Cote et al. | Dec 2014 | B2 |
9113118 | Silverstein et al. | Aug 2015 | B2 |
9131196 | Lim et al. | Sep 2015 | B2 |
9131644 | Osborne | Sep 2015 | B2 |
9195891 | Scharf et al. | Nov 2015 | B2 |
9230177 | Dolinar et al. | Jan 2016 | B2 |
9342858 | Cote et al. | May 2016 | B2 |
20050100220 | Keaton et al. | May 2005 | A1 |
20100079333 | Janky et al. | Apr 2010 | A1 |
20110276336 | Sweely | Nov 2011 | A1 |
20130066666 | Anderson et al. | Mar 2013 | A1 |
20130124239 | Rosa et al. | May 2013 | A1 |
20130321677 | Cote et al. | Dec 2013 | A1 |
20130322753 | Lim et al. | Dec 2013 | A1 |
20140032271 | Nordstrand | Jan 2014 | A1 |
20140067270 | Gail | Mar 2014 | A1 |
20140358486 | Osborne | Dec 2014 | A1 |
20140379228 | Batcheller et al. | Dec 2014 | A1 |
20150015692 | Smart | Jan 2015 | A1 |
20150116330 | Chiocco et al. | Apr 2015 | A1 |
20150206255 | Groeneveld | Jul 2015 | A1 |
20150327440 | Dybro et al. | Nov 2015 | A1 |
20150370935 | Starr | Dec 2015 | A1 |
20150371161 | Mueller et al. | Dec 2015 | A1 |
20160050840 | Sauder et al. | Feb 2016 | A1 |
20160055593 | Groeneveld | Feb 2016 | A1 |
20160063639 | Groeneveld | Mar 2016 | A1 |
20160066505 | Bakke et al. | Mar 2016 | A1 |
20160071223 | Rupp et al. | Mar 2016 | A1 |
20160071410 | Rupp et al. | Mar 2016 | A1 |
20160073573 | Ethington et al. | Mar 2016 | A1 |
20160078304 | Bremer et al. | Mar 2016 | A1 |
20160078570 | Ethington et al. | Mar 2016 | A1 |
20160084813 | Anderson et al. | Mar 2016 | A1 |
20160084987 | Dybro et al. | Mar 2016 | A1 |
20160171680 | Lobell | Jun 2016 | A1 |
20160180473 | Groeneveld | Jun 2016 | A1 |
20160202227 | Mathur et al. | Jul 2016 | A1 |
20160232621 | Ethington et al. | Aug 2016 | A1 |
20160247082 | Stehling et al. | Aug 2016 | A1 |
20160259089 | McKinnon | Sep 2016 | A1 |
20160290918 | Xu et al. | Oct 2016 | A1 |
20180188417 | Spatzier | Jul 2018 | A1 |
Number | Date | Country |
---|---|---|
2663917 | Oct 2010 | CA |
2697608 | Sep 2011 | CA |
104635281 | May 2015 | CN |
107300703 | Oct 2017 | CN |
108375807 | Aug 2018 | CN |
2491721 | Jun 2016 | EP |
2531908 | May 2016 | GB |
WO 2010126783 | Nov 2010 | WO |
WO 2010126783 | Nov 2010 | WO |
WO 2015092800 | Jun 2015 | WO |
WO 2015200489 | Dec 2015 | WO |
WO 2016037105 | Mar 2016 | WO |
WO 2016099723 | Jun 2016 | WO |
WO 2016127094 | Aug 2016 | WO |
WO 2016144818 | Sep 2016 | WO |
WO 2016160384 | Oct 2016 | WO |
WO 2016164147 | Oct 2016 | WO |
WO 2017104882 | Jun 2017 | WO |
Entry |
---|
European Commission—Joint Research Centre, “Interpolation of observed weather,” Apr. 2018, 7 pages [Online] [Retrieved Nov. 28, 2018], Retrieved from the internet: <URL https://marswiki.jrc.ec.europa.eu/agri4castwiki/index.php/Interpolation of observed weather>. |
PCT International Search Report and Written Opinion, PCT Application No. PCT/CA2018/051127, dated Sep. 12, 2018, nine pages. |
Stahl, K. et al., “Comparison of approaches for spatial interpolation of daily air temperature in a large region with complex topography and highly variable station density,” Agricultural and Forest Meteorology, vol. 139, No. 3-4, 2006, pp. 224-236. |
Van Der Voet, P. et al., “Spatial interpolation of daily meteorological data,” DLO Winand Staring Centre, vol. 53, No. 3, 1994, 100 pages. |
Sluiter, R., “Interpolation Methods for Climate Data”, Intern Rapport; IR Apr. 2009, Literature review, KNMI, R&D Information and Observation Technology, De Bilt, Nov. 19, 2008, Version 1.0, 28 pages. |
Hofstra, N. et al., “Comparison of Six Methods for the Interpolation of Daily, European Climate Data,” American Geophysical Union, Journal of Geophysical Research, vol. 113, D21110, 2008, 19 pages. |
Li, J. et al., “A Review of Spatial Interpolation Methods for Environmental Scientists,” Australian Government, Record 2008/23, Geocat#68229, 2008, 154 pages. |
Number | Date | Country | |
---|---|---|---|
20190079214 A1 | Mar 2019 | US |
Number | Date | Country | |
---|---|---|---|
62558643 | Sep 2017 | US |