COMPUTER-IMPLEMENTED METHODS AND NON-TRANSITORY COMPUTER-READABLE STORAGE MEDIUM FOR IMPROVING ICE COVERAGE PREDICTIONS USING MACHINE LEARNING, SATELLITE, THERMODYNAMIC, AND IN SITU OBSERVATIONS

Abstract

To provide a computing device, non-transitory computer-readable storage medium and method for performing predictions with machine learning to determine ice coverage.SOLUTION: A computer implemented method of predicting ice coverage on a body of water includes generating a first ice cover prediction with a thermodynamics module 104, and generating a second ice cover prediction with a machine learning module 106. The first ice cover prediction is combined with the second ice cover prediction to generate a combined ice cover prediction. Error statistics are computed based on a comparison of the combined ice cover prediction with an ice coverage observation result, and the combined ice cover prediction is updated based on the error statistics.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates generally to artificial intelligence and machine learning systems, and more particularly, to methods and systems for predicting ice coverage on a body of water. [Background technology]

[0002] Many lakes and parts of the oceans are covered with ice in winter. Ice acts as a buffer between the atmosphere and the water, strongly influencing the flux of heat and momentum, the exchange of gases, and the penetration of light. This has profound effects on the movement of water and nutrients and its biological evolution. A reliable prediction of ice cover can ensure a more accurate simulation of the fluid dynamics, along with nutrients, chemistry, and biology during the ice-covered months. Summary of the Invention [Problem to be solved by the invention]

[0003] However, ice cover is often partial, i.e., part of the body of water is covered with ice and the rest is open water. Partial ice cover can exist for days, weeks or even months, and this ice cover expands and contracts as the weather changes. During such periods, some parts of the body of water are exposed to the atmosphere, while other parts of the body of water have this ice cover buffer. This has a local effect on the water cycle, and also a more general effect if partial ice cover exists for many days.

[0004] Currently, hydrodynamic models are used to simulate water movement. To accurately simulate the long-term hydrodynamics of partial ice cover, the hydrodynamic models require knowledge of the spatial extent of the ice cover.

[0005] According to various embodiments, computing devices, non-transitory computer-readable storage media, and methods are provided for performing machine learning predictions to determine ice coverage using at least one of satellite data, thermodynamics, and in situ observation systems. [Means for solving the problem]

[0006] In one embodiment, a computer-implemented method for predicting ice coverage on a body of water includes generating a first ice coverage forecast using a thermodynamics module and generating a second ice coverage forecast using a machine learning module. The first ice coverage forecast is combined with the second ice coverage forecast to generate a combined ice coverage forecast. An error statistic is calculated based on a comparison of the combined ice coverage forecast to ice coverage observations, and the combined ice coverage forecast is updated based on the error statistic.

[0007] In some embodiments, the thermodynamic module includes a hydrodynamic model and thermodynamic equations used to generate the first ice cover prediction. In some embodiments, ice cover in the hydrodynamic model is updated based on the error statistics.

[0008] In some embodiments, the machine learning module includes a module for generating a trained model and a predicted local weather condition. In some embodiments, a feedback loop is provided to the trained model based on error statistics.

[0009] In some embodiments, the method further includes applying weights and thresholds to the first ice cover forecast and the second ice cover forecast to generate a combined ice cover forecast. In some embodiments, the weights and thresholds are applied to each grid cell defining a surface area of ​​the body of water. In some embodiments, the weight values ​​and thresholds for each grid cell are defined individually based on a predetermined period throughout the year.

[0010] In some embodiments, a tree-based model is used to optimize the combined ice cover predictions.

[0011] In some embodiments, the method further includes making direct observations of ice coverage on the body of water with an observation device. In some embodiments, a path of travel for the observation device is optimized based on the combined ice coverage forecasts and observations of whether ice is present in grid cells defining a surface area of ​​the body of water. In some embodiments, the observation device is an in-situ automated observation device.

[0012] According to various embodiments, a computer-implemented method for predicting ice coverage on a body of water includes generating a first ice coverage forecast using a thermodynamics module and generating a second ice coverage forecast using a machine learning module. The first ice coverage forecast is combined with the second ice coverage forecast to generate a combined ice coverage forecast based on an initial set of weight values ​​and threshold values. An error statistic is calculated based on a comparison of the combined ice coverage forecast to ice coverage observations from an automated observation device, and the weight values ​​and threshold values ​​are updated based on the error statistic. An updated combined ice coverage forecast is generated using the updated weight values ​​and threshold values.

[0013] The concepts discussed herein provide systems and methods that improve upon approaches currently used to determine ice coverage. The systems and methods discussed herein can improve the accuracy of ice coverage predictions, including prediction of ice coverage expansion and contraction. As discussed below, the systems and methods can further improve predictions through the use of in-situ observation systems, thus including the potential to confirm predictions from models and improve the models themselves.

[0014] These and other features will become apparent from the following detailed description of illustrative embodiments thereof, which is to be read in connection with the accompanying drawings. [Brief description of the drawings]

[0015] [Figure 1] 1 is a representation of various components used in a system for predicting ice coverage on a body of water, according to one embodiment of the present disclosure. [Diagram 2] A pictorial representation of a water body divided into a grid, where signs of ice cover are predicted using a combined hydrodynamic and machine learning model. [Diagram 3] 3 is a pictorial representation illustrating how the fluid dynamics model and the machine learning model of FIG. 2 are combined to generate a combined prediction, consistent with an illustrative embodiment. [Figure 4] 4 is a pictorial representation showing the combined forecast of FIG. 3 and further illustrating how in-situ autonomous observations can be used to update the combined forecast, consistent with certain illustrative embodiments; [Diagram 5] 1 is a flowchart illustrating acts involved in generating and updating ice coverage forecasts consistent with certain example embodiments. [Figure 6] Figure 2 is a functional block diagram of a computer hardware platform that can be used to implement the ice coverage forecasting engine. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] The drawings are of exemplary embodiments. They do not depict every embodiment. Other embodiments may be used in addition or instead. Details that may be obvious or unnecessary may be omitted to save space or for a more effective illustration. Some embodiments may be practiced with additional components or steps and / or without all components or steps shown. When the same numeral appears in different drawings, it refers to the same or similar components or steps.

[0017] In the following detailed description, numerous specific details are presented by way of example to provide a thorough understanding of the relevant teachings. However, it should be apparent that the present teachings may be practiced without such details. In other instances, well-known methods, procedures, components, or circuit elements, or combinations thereof, have been described at a relatively high level, without detail, in order to avoid unnecessarily obscuring aspects of the present teachings.

[0018] Unless specifically noted otherwise, as will be apparent from the description which follows and the claims, it is understood that throughout the specification, descriptions utilizing terms such as "processing," "computing," "calculating," "determining," or the like refer to operations and / or processes of a computer or computing system, or similar electronic computing device, that manipulate and / or transform data represented as physical, e.g., electronic, quantities in the registers or memory of the computing system into other data similarly represented as physical quantities in the memory, registers or other such information storage, transmission or display device of the computing system.

[0019] As discussed in more detail below, the present disclosure generally relates to a system for predicting ice cover for a body of water by using a mixed model approach that utilizes both physically based and machine learning models to predict realistic ice cover. The machine learning model can be enhanced with observations. The model can include feedback, as described in more detail below, to improve the predictions over time. The present disclosure further relates to a method for simulating hydrodynamics during periods of ice cover by combining predictions from a machine learning model trained on satellite-derived ice cover concentrations with predictions from a thermodynamic-based ice model driven by a hydrodynamic model to determine ice concentration and flag grid cells in the hydrodynamics as having ice cover.

[0020] There are existing methods for predicting ice concentration using satellite observations and weather data using machine learning. Additionally, there are existing methods for predicting ice formation in hydrodynamic models via numerical discretization of water thermodynamics using differential equations. The method and system of the present invention uses a combined optimized feedback of both machine learning and hydrodynamic models, as discussed below, to improve simulations using hydrodynamic models.

[0021] Reference is now made in detail to the details illustrated in the accompanying drawings and discussed below.

[0022] 1 and 2, a body of water 100 may include ice cover 102, indicated by a diagonal line on the body of water 100. A thermodynamic module 104 of the ice prediction engine 106 may receive weather data from a weather station 112 and may use this data in a hydrodynamic model 114 in conjunction with a thermodynamic equation 116 to make ice cover predictions. A machine learning module 108 may use satellite data from a satellite 110 to correlate observed ice concentration and local climate to make predictions of future ice cover. The thermodynamic module 104 may perform calculations every time step of the hydrodynamic model 114, for example, up to every 3 seconds, providing the thermodynamic module 104 with high temporal resolution. Information for ice cover predictions may be produced with high spatial resolution.

[0023] The machine learning module 108 can use correlations between observed ice concentrations and local climate to make predictions. The machine learning module 108 calculations can occur when new weather information is available and can be gated by the selection of feature values ​​used to train the model, for example, using mean daily temperature. In some embodiments, the trained model 202 of the machine learning module 108 can generate predicted local weather conditions 204 to generate an ice cover prediction 230.

[0024] 2 and 3, the ice prediction engine 106 may generate ice coverage predictions 220, 230 for each of the thermodynamic module 104 and the machine learning module 108. These ice coverage predictions 220, 230 may be combined, as discussed in more detail below, to generate a combined prediction 240. The combined prediction 240 may illustrate the extent of ice coverage over the body of water 100, indicated by grid portions marked with x's.

[0025] 4, in act 400, the combined prediction 240 can be compared to observations made on actual ice cover and error statistics can be calculated. Thresholds and parameters, as described below, used to generate the combined prediction 240 (FIG. 3) from the predictions 220, 230 from the thermodynamics module 104 and the machine learning module 108 can be adjusted in act 402 based on the calculated error statistics. The combined prediction 240 can be updated in act 404.

[0026] In some embodiments, an in-situ autonomous observation device 120 (see FIG. 1 , also referred to as observation device 120) can be used to observe actual ice coverage in act 406, and such observations can be used in act 400 to compare these observations with the combined prediction 240 and calculate error statistics.

[0027] The observation device 120 can be used to complement remotely sensed data, such as from satellites 110. The observation device 120 can take a variety of forms, such as an underwater vehicle (UAV), a surface floating device, such as a floating dock, an aircraft, such as a drone, a snowmobile, an amphibious vehicle, or the like. The observation device 120 can be equipped with sensors to detect the presence of ice coverage at any location. In some embodiments, the observation device 120 can be optimized to explore areas based on the combined forecast 240 and selectively explore those areas that border ice coverage, which can be advantageous for larger bodies of water where exploring the entire body of water may not be practical or achievable within a given time. In some embodiments, the observation device 120 can include machine learning algorithms that provide the ability to use weather data in conjunction with the combined forecast 240 to determine a desired search space.

[0028] With the above overview of the example ice forecast engine 106 (see FIG. 2 ), it may be useful to next provide a high-level discussion of an example process. To that end, FIG. 5 presents an example process for generating ice coverage forecasts through both thermodynamic and machine learning modules. The process 500 is illustrated as a collection of blocks in a logical flow chart, which represent a sequence of operations that may be implemented in hardware, software, or a combination thereof. In the context of software, these blocks represent computer-executable instructions that perform the recited operations when executed by one or more processors. In general, computer-executable instructions may include routines, programs, objects, components, data structures, and the like that perform functions or implement abstract data types. In each process, the order in which the operations are described is not intended to be construed as a limitation, and any number of the blocks described may be combined in any order and / or in parallel to implement the process. For discussion purposes, the process 500 is described with reference to the ice forecast engine 106 of FIG. 2 .

[0029] 5, a process 500 for predicting ice coverage of a body of water may use the thermodynamics module 104 to generate an ice / no-ice prediction at each grid cell for each time step in act 502. The process 500 may use the machine learning module 108 to generate an ice / no-ice prediction at each grid cell using, for example, a deep neural net in act 504.

[0030] In act 506, weights and thresholds may be applied to combine the predictions of acts 502 and 504 to generate a combined prediction in act 508. The problem of applying weights and thresholds in act 506 may be fitted to a tree-based model (e.g., Random Forest, XGBoost, or the like) to optimize the prediction, although other algorithmic approaches could be used. Random weight values ​​and thresholds may be selected to provide a first prediction. The tree-based model may be trained using the observations, and weight values ​​and thresholds for each grid cell may be adjusted based on error statistics in act 510. For example, weight values ​​and thresholds may be adjusted based on the L2 norm in act 512 to update the prediction in act 514. In some embodiments, hyperparameter tuning of the tree-based model (e.g., adjusting the number of branches) may be performed for optimized performance of the tree-based model to adjust the weight values ​​and thresholds in act 512. Adjusted weights and thresholds can be defined for a given grid cell of a surface area of ​​a body of water, such as the grid cell shown in the combined ice cover forecast 240 in FIG. 3. Subsequent training can occur as more observations become available. In some embodiments, weights and thresholds for a given grid cell may be defined for a given time of year, and since the predictive skill of ice cover from hydrodynamic models and satellite-based machine learning models may vary seasonally, different weights and thresholds may be provided, for example, for each month.

[0031] When a combined forecast is generated (in acts 506 and 508) or updated (in act 514), the updated ice coverage forecast can be fed back to the thermodynamics module 104 in act 516 for use in the fluid dynamics model 114. In some embodiments, based on the error statistics determined in act 510, the ice prediction engine 106 (see FIG. 2 ) updates the trained model of the machine learning module 108 so that the forecast made in act 504 is more accurate.

[0032] In some embodiments, process 500 may use an in situ automated observation device, and in act 518 may determine an optimal path of travel for the observation device, and in act 520 may use the observation device to observe whether ice is present at a given location on the body of water. The optimal path of travel may be adjusted to precisely define the boundary between ice coverage and non-ice on the body of water.

[0033] In some embodiments, the trained model 202 can be trained using hindcasts performed by the fluid dynamics model 114. In other words, the output from the fluid dynamics model 114 can be used as feature values ​​in the machine learning module 108 to generate links between the two modules 104, 108.

[0034] In some embodiments, the ice coverage prediction engine 106 may be used for iceberg tracking. For example, given local weather information (feature values), the machine learning module 108 can be trained on possible iceberg trajectories using satellite-derived trajectories (target values). Predictions of iceberg trajectories can be obtained from combining hydrodynamic models with the machine learning module in a manner similar to that described above.

[0035] 6 provides a functional block diagram of a computer hardware platform 600 that may be used to implement a specially configured computing device capable of hosting an ice coverage prediction engine 106. The ice coverage prediction engine 106 may include a thermodynamics module 104 and a machine learning module 108, as previously discussed. The thermodynamics module 104 may include a fluid dynamics model 114 and a thermodynamics equation 116, as previously discussed. The machine learning module 108 may include a trained model 202 and a predicted weather condition 204, as previously discussed. In particular, FIG. 6 illustrates a network or host computer platform 600 that may be used to implement a suitably configured engine 106, such as that illustrated in FIG. 2.

[0036] Computer platform 600 may include a central processing unit (CPU) 610, a hard disk drive (HDD) 620, random access memory (RAM) and / or read only memory (ROM) 630, a keyboard 650, a mouse 660, a display 670, and a communication interface 680 connected to a system bus 640.

[0037] In one embodiment, HDD 620 has capabilities including storing programs capable of executing various processes, such as ice coverage prediction engine 106, in the manner described herein.

[0038] In one embodiment, the communications interface 680 may be configured to receive data from a satellite 110, a weather station 112, an observation device 120, or the like, as illustrated in FIG.

[0039] conclusion The description of various embodiments of the present teachings has been presented for illustrative purposes, but they are not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terms used in this specification have been selected to best explain the principles of the embodiments, practical applications, or technical improvements over the art found in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

[0040] While the above describes what is believed to be the best mode or other examples or combinations thereof, it is understood that various modifications may be made herein, that the subject matter disclosed herein may be implemented in various forms or examples, and that the present teachings may be applied to many applications, only a few of which have been described herein. It is intended by the appended claims to claim all such applications, modifications and variations that are within the true scope of the present teachings.

[0041] The components, steps, features, objects, benefits and advantages discussed herein are merely illustrative. No discussion of any of them or their related aspects is intended to limit the scope of protection. Although various advantages have been discussed herein, it will be understood that not all embodiments necessarily include all advantages. Unless otherwise specified, all measurements, values, ratings, positions, dimensions, sizes, and other specifications described herein, including the appended claims, are approximate and not precise. They are intended to have a reasonable range consistent with the function to which they relate and with what is customary in the technical field to which they pertain.

[0042] Many other embodiments are contemplated, including embodiments having fewer, additional, or different or different combinations of components, steps, features, objects, benefits, and advantages, including embodiments in which the components or steps, or both, are arranged or ordered differently or in different combinations.

[0043] Aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer readable program instructions.

[0044] These computer readable program instructions may be provided to a suitably configured computer, special purpose computer, or other programmable data processing apparatus to produce a machine such that the instructions, which execute via a processor of the computer or other programmable data processing apparatus, produce means for implementing the functions / acts specified in one or more blocks of the flowcharts and / or block diagrams. These computer readable program instructions may be stored in a computer readable storage medium, and the program instructions may direct a computer, programmable data processing apparatus, or other device, or combination thereof, to function in a manner such that the computer readable storage medium having instructions stored therein comprises an article of manufacture containing instructions that implement aspects of the functions / acts specified in one or more blocks of the flowcharts and / or block diagrams.

[0045] The computer readable program instructions may be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus, or other device to produce a computer-implemented process, such that the instructions executing on the computer, other programmable apparatus, or other device implement the functions / acts specified in one or more blocks in the flowchart and / or block diagram.

[0046] The call flows, flowcharts, and block diagrams in the figures herein illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowcharts or block diagrams may represent a module, segment, or portion of instructions, comprising one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the blocks may occur in an order other than that noted in the figures. For example, two blocks shown in succession may in fact be executed substantially simultaneously, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. It will also be noted that each block in the block diagrams and / or flowchart illustrations, as well as combinations of blocks in the block diagrams and / or flowchart illustrations, may be implemented by a dedicated hardware-based system that performs the specified functions or acts, or that executes a combination of dedicated hardware and computer instructions.

[0047] Although the above has been described in conjunction with exemplary embodiments, it is understood that the term "exemplary" means merely an example, rather than best or optimal. Except as expressly stated immediately above, nothing described or illustrated is intended, and should not be construed, to result in the dedication to the public of any component, step, feature, object, benefit, advantage, or equivalent, whether or not it is recited in the claims.

[0048] It will be understood that the terms and expressions used herein have the ordinary meanings given to such terms and expressions with respect to their corresponding respective fields of inquiry and study, unless a specific meaning is otherwise presented herein. Relative terms such as first and second and the like may be used solely to distinguish one entity or operation from another, without necessarily requiring or implying any actual relationship or order between such entities or operations. The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus comprising a list of elements may include not only those elements, but other elements not expressly listed or inherent in such process, method, article, or apparatus. An element preceded by "a" or "an" does not, without further constraints, preclude the presence of additional similar elements in the process, method, article, or apparatus comprising that element.

[0049] The Abstract of the Disclosure is provided to allow the reader to quickly ascertain the nature of the technical disclosure. The disclosure is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the claims. In addition, in the above Detailed Description, it can be seen that various features have been grouped into various embodiments for the purpose of streamlining the disclosure. This method of disclosure should not be interpreted as reflecting an intention that the claimed embodiments have more features than are expressly recited in each claim. Rather, as the appended claims reflect, inventive subject matter lies in less than all features of a single disclosed embodiment. Thus, the appended claims are hereby incorporated into the Detailed Description, with each claim standing on its own as separately claimed subject matter. [Explanation of symbols]

[0050] 100 bodies of water 102 Ice Cover 104 Thermodynamics Module 106 Ice Prediction Engine 108 Machine Learning Module 110 satellite 114 Fluid Dynamics Model 116 Thermodynamic equations 120 Observation Device

Claims

1. 1. A computer-implemented method for predicting ice coverage on a body of water, comprising: generating a first ice cover forecast using a thermodynamics module; generating a second ice cover forecast using a machine learning module; combining the first ice coverage forecast and the second ice coverage forecast to generate a combined ice coverage forecast; calculating an error statistic based on a comparison of the combined ice coverage prediction and ice coverage observations; updating the combined ice coverage forecast based on the error statistics; A method comprising:

2. The computer-implemented method of claim 1 , wherein the thermodynamics module includes a hydrodynamic model and one or more thermodynamic equations used to generate the first ice cover forecast.

3. The computer-implemented method of claim 2 , further comprising updating the fluid dynamics model based on the error statistics.

4. The computer-implemented method of claim 1 , wherein the machine learning module includes a module for generating a trained model and predicted local weather conditions.

5. 5. The computer-implemented method of claim 4, further comprising providing a feedback loop to the trained model based on the error statistics.

6. The computer-implemented method of claim 1 , further comprising applying weights and thresholds to generate the combined ice coverage forecast from the first ice coverage forecast and the second ice coverage forecast.

7. The computer-implemented method of claim 6 , wherein the weighting values ​​and thresholds are applied to each grid cell defining a surface area of ​​the body of water.

8. The computer-implemented method of claim 6 , wherein the weight values ​​and thresholds for each grid cell are defined individually based on predetermined periods throughout the year.

9. The computer-implemented method of claim 6 , wherein a tree-based model is used to optimize the combined ice cover forecast.

10. The computer-implemented method of claim 1 , further comprising making direct observations of ice coverage on the body of water with an observation device.

11. 11. The computer-implemented method of claim 10, wherein a path of travel for the observation device is optimized based on the combined ice coverage prediction and an observation of whether ice is present in grid cells defining a surface area of ​​the body of water.

12. 12. The computer-implemented method of claim 11, further comprising updating the heading path when a discrepancy occurs regarding ice coverage based on the combined ice coverage forecast and the observations of the observation device.

13. The computer-implemented method of claim 10 , wherein the observation device is an in-situ automatic observation device.

14. 1. A computer-implemented method for predicting ice coverage on a body of water, comprising: generating a first ice cover forecast using a thermodynamics module; generating a second ice cover forecast using a machine learning module; combining the first ice coverage forecast and the second ice coverage forecast to generate a combined ice coverage forecast based on an initial set of weight values ​​and threshold values; calculating error statistics based on a comparison of the combined ice coverage forecast and ice coverage observations from an automated observation device; updating the weights and thresholds based on the error statistics; using the updated weights and thresholds to generate an updated combined ice coverage forecast; A method comprising:

15. 15. The computer-implemented method of claim 14, further comprising updating a path of travel of the automated observation device when a discrepancy occurs regarding ice coverage based on the combined ice coverage forecast and the observations made of ice coverage using the automated observation device.

16. The computer-implemented method of claim 14 , wherein the weighting values ​​and thresholds are applied to each grid cell defining a surface area of ​​the body of water.

17. 15. The computer-implemented method of claim 14, wherein the weight values ​​and thresholds for each grid cell are defined individually based on predetermined periods throughout the year.

18. The computer-implemented method of claim 14 , wherein a tree-based model is used to optimize the combined ice cover prediction.

19. A non-transitory computer readable storage medium tangibly embodying computer readable medium program code having computer readable instructions that, when executed, cause a computing device to perform a method for forecasting ice coverage on a body of water, the method comprising: generating a first ice cover forecast using a thermodynamics module; generating a second ice cover forecast using a machine learning module; combining the first ice coverage forecast and the second ice coverage forecast to generate a combined ice coverage forecast; calculating an error statistic based on a comparison of the combined ice coverage prediction and ice coverage observations; updating the combined ice coverage forecast based on the error statistics; 1. A non-transitory computer readable storage medium comprising:

20. The execution of the code by a processor includes: generating the combined ice coverage forecast by applying weights and thresholds to the first ice coverage forecast and the second ice coverage forecast; 20. The non-transitory computer-readable storage medium of claim 19, further configuring the computing device to perform actions including:

21. The execution of the code by a processor includes: making direct observations of ice coverage in said body of water using an automated in-situ observation device.

20. The non-transitory computer-readable storage medium of claim 19, further configuring the computing device to perform actions including:

22. A non-transitory computer readable storage medium tangibly embodying computer readable program code having computer readable instructions that, when executed, cause a computing device to perform a method for forecasting ice coverage on a body of water, the method comprising: generating a first ice cover forecast using a thermodynamics module; generating a second ice cover forecast using a machine learning module; combining the first ice coverage forecast and the second ice coverage forecast to generate a combined ice coverage forecast based on an initial set of weight values ​​and threshold values; calculating error statistics based on a comparison of the combined ice coverage forecast and ice coverage observations from an automated observation device; updating the weights and thresholds based on the error statistics; using the updated weights and thresholds to generate an updated combined ice coverage forecast; 1. A non-transitory computer readable storage medium comprising:

23. The execution of the code by a processor includes: updating a travel path of the automated observation device when a discrepancy occurs regarding ice coverage based on the combined ice coverage prediction and the observations made of ice coverage using the automated observation device.

23. The non-transitory computer-readable storage medium of claim 22, further configuring the computing device to perform actions including:

24. 23. The non-transitory computer-readable storage medium of claim 22, wherein the weighting values ​​and thresholds are applied to each grid cell defining a surface area of ​​the body of water.

25. 25. The non-transitory computer-readable storage medium of claim 24, wherein the weight values ​​and thresholds for each grid cell are defined individually based on predetermined periods throughout the year.

Images