Estimating confidence bounds for rainfall adjustment values

Patent metadata

Field	Value
Publication number	US-10467540-B2
Application number	US-201615171471-A
Country	US
Kind code	B2
Filing date	Jun 2, 2016
Priority date	Jun 2, 2016
Publication date	Nov 5, 2019
Grant date	Nov 5, 2019

Abstract

A method for estimating confidence bounds for adjusted rainfall values for a set of geo-locations using agricultural data comprises using a server computer system that receives, via a network, agricultural data records that are used to estimate rainfall values for the set of geo-locations. Within the server computer system, rainfall calculation instructions receive digital data including observed radar and rain-gauge agricultural data records. The computer system then aggregates the agricultural data records and creates and stores the agricultural data sets. The agricultural data records are then used to estimate adjusted rainfall values for a set of geo-locations. Rainfall confidence bounds instructions estimate a set of confidence bounds for each of the adjusted rainfall values for the set of geo-locations. The set of confidence bounds provide a range for each of the adjusted rainfall values that represents a particular level of confidence associated with each of the adjusted rainfall values.

First claim

What is claimed is:

1. A method comprising:
using rainfall calculation instructions in a server computer system, receiving one or more agricultural data records that represent observed agricultural data points for specific geo-locations at a specific time, wherein the observed agricultural data points include at least one of observed radar data or rain gauge data that comprise a plurality of precipitation data values and that have been obtained from radar computers coupled to radars or a weather station computers coupled to rain gauges;
using the rainfall calculation instructions, aggregating the one or more agricultural data records into one or more agricultural data sets, where each agricultural data set from the one or more agricultural data sets represents a single type of observed agricultural data;
using rainfall estimation instructions in the server computer system, estimating a set of adjusted rainfall values for a set of geo-locations using the one or more agricultural data sets, wherein the set of adjusted rainfall values represent calculated rainfall value estimations;
using rainfall confidence bounds instructions, estimating a set of confidence bounds for the set of adjusted rainfall values, wherein the set of confidence bounds provides a range for each adjusted rainfall value in the set of adjusted rainfall values that represents a particular level of confidence for the set of adjusted rainfall values;
using confidence bound aggregation instructions in the server computer system, determining an aggregated confidence bounds for a rainfall estimate over a specific period of time by calculating lower precipitation estimates by aggregating lower confidence bounds from the set of confidence bounds for each geo-location over the specific period of time; calculating upper precipitation estimates by aggregating upper confidence bounds from the set of confidence bounds for each geo-location over the specific period of time; using the confidence bound aggregation instructions, generating a set of aggregated confidence bounds comprising aggregated lower and aggregated upper confidence bounds for each geo-location in the set of geo-locations.

2. The method of claim 1 , wherein estimating a set of adjusted rainfall values for a set of geo-locations using the one or more agricultural data sets further comprises:
transforming the one or more agricultural data sets into one or more agricultural distribution sets, where the one or more agricultural distribution sets represent a normalized distribution of the one or more agricultural data sets;
generating a covariate matrix from the one or more agricultural distribution sets and storing the covariate matrix in digital memory, by deriving at least some values within the covariate matrix from the one or more agricultural distribution sets;
estimating the set of adjusted rainfall values for the set of geo-locations that correspond to the values within the covariate matrix using defined regression parameters from a digital rainfall regression model and previously stored adjusted rainfall data sets.

3. The method of claim 2 , wherein estimating the set of adjusted rainfall values for the set of geo-locations that correspond to the values within the covariate matrix using defined regression parameters from a digital rainfall regression model and previously stored adjusted rainfall data sets comprises generating a joint distribution between the one or more agricultural distribution sets and the previously stored adjusted rainfall data sets.

4. The method of claim 3 , wherein estimating a set of confidence bounds for the set of adjusted rainfall values further comprises:
generating an inverse cumulative distribution based upon mean and variance values for the set of adjusted rainfall values from the joint distribution;
determining a lower confidence bound as a 10th percentile of the inverse cumulative distribution;
determining an upper confidence bound as a 90th percentile of the inverse cumulative distribution;
compiling the set of confidence bounds for the set of adjusted rainfall values, wherein each confidence bounds in the set of confidence bounds includes a pair of bounds comprising the lower confidence bound and the upper confidence bound.

5. The method of claim 1 , wherein estimating a set of adjusted rainfall values for a set of geo-locations using the one or more agricultural data sets further comprises:
calculating a correction factor that describes correlation of values between agricultural data points that represent observed radar data and agricultural data points that represent rain gauge data;
wherein the correction factor is an aggregate ratio calculated as a sum of the agricultural data points that represent rain gauge data from the one or more agricultural data sets divided by a sum of the agricultural data points that represent observed radar data from the one or more agricultural data sets;
calculating the set of adjusted rainfall values for the set of geo-locations by multiplying agricultural data points that represent observed radar data for the set of geo-locations by the correction factor.

6. The method of claim 5 , wherein estimating a set of confidence bounds for the set of adjusted rainfall values further comprises:
calculating a set of subsampled correction factors from the one or more agricultural data sets, wherein each subsampled correction factor of the set of subsampled correction factors is calculated from a subset of the one or more agricultural data sets that represent agricultural data records received at over a specific time period;
generating a correction factor distribution of the set of subsampled correction factors;
generating an inverse cumulative distribution function based upon the correction factor distribution of the set of subsampled correction factors;
determining a lower confidence bound as a 10th percentile of the inverse cumulative distribution function;
determining an upper confidence bound as a 90th percentile of the inverse cumulative distribution function;
compiling the set of confidence bounds for the set of adjusted rainfall values , wherein each confidence bounds in the set of confidence bounds includes a pair of bounds comprising the lower confidence bound and the upper confidence bound.

7. The method of claim 5 , wherein estimating a set of confidence bounds for the set of adjusted rainfall values further comprises:
calculating a set of correction factors based upon the one or more agricultural data sets, wherein each correction factor in the set of correction factors represents a specific geo-location within the one or more agricultural data sets and is a ratio calculated as rain gauge data at the specific geo-location divided by observed radar data at the specific geo-location;
generating a correction factor distribution set based upon the set of correction factors;
determining a lower confidence bound as a 10th percentile of the correction factor distribution set;
determining an upper confidence bound as a 90th percentile of the correction factor distribution set;
compiling the set of confidence bounds for the set of adjusted rainfall values, wherein each confidence bounds in the set of confidence bounds includes a pair of bounds comprising the lower confidence bound and the upper confidence bound.

8. The method of claim 1 , wherein estimating a set of adjusted rainfall values for a set of geo-locations using the one or more agricultural data sets further comprises:
generating a gamma distribution set based upon the one or more agricultural data sets, wherein the gamma distribution set comprises rain gauge-to-radar ratio values that are each calculated as rain gauge data divided by corresponding observed radar data for a specific ge

Assignees

• Climate Corp

Inventors

• Leeds Bill
• Lakshmanan Valliappa
• Alvarez Francisco
• Hryniw Natalia

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N3/126

  Evolutionary algorithms, e.g. genetic algorithms or genetic programming · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N3/006

  based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] · CPC title

• G01S7/003

  Transmission of data between radar, sonar or lidar systems and remote stations · CPC title