US 12,031,964 B2
Enhancing spatial and temporal resolution of greenhouse gas emission estimates for agricultural fields using cohort analysis techniques
Isaac Waweru Wambugu, Nairobi (KE); Ranjini Bangalore Guruprasad, Bangalore (IN); Manikandan Padmanaban, Bangalore (IN); Kumar Saurav, Bangalore (IN); Ivan Kayongo, Nairobi (KE); and Jagabondhu Hazra, Bangalore (IN)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Oct. 6, 2021, as Appl. No. 17/495,582.
Prior Publication US 2023/0106473 A1, Apr. 6, 2023
Int. Cl. G01N 33/00 (2006.01)
CPC G01N 33/0036 (2013.01) 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
obtaining non-greenhouse gas remote sensing data pertaining to at least a portion of multiple agricultural fields, and contextual information pertaining to at least a portion of the multiple agricultural fields;
determining one or more cohorts among the multiple agricultural fields based at least in part on deriving one or more agricultural field-specific features from at least a portion of the obtained non-greenhouse gas remote sensing data and at least a portion of the obtained contextual information;
computing at least one agricultural field-level time series of greenhouse gas emission estimates for each of the one or more cohorts by processing at least a portion of the obtained non-greenhouse gas remote sensing data and at least a portion of the obtained contextual information using at least one process-based model;
calculating at least one bias correction for each of the one or more cohorts by processing, using a time series learning model, the at least one agricultural field-level time series of greenhouse gas emission estimates and one or more background greenhouse gas emission estimates pertaining to at least a portion of the one or more cohorts;
generating at least one updated greenhouse gas emission estimate for at least a portion of the one or more cohorts based at least in part on at least one initial greenhouse gas emission estimate derived from greenhouse gas remote sensing data pertaining to at least a portion of the multiple agricultural fields and the at least one calculated bias correction for the at least a portion of the one or more cohorts, wherein the at least one updated greenhouse gas emission estimate comprises enhanced spatial resolution and temporal resolution as compared to the at least one initial greenhouse gas emission estimate; and
performing one or more automated actions based at least in part on the at least one updated greenhouse gas emission estimate, wherein performing one or more automated actions comprises automatically training at least a portion of the time series learning model using at least a portion of the at least one updated greenhouse gas emission estimate;
wherein the method is carried out by at least one computing device.