Milk and Cattle (in weight) Productivity Estimation | ||||||||
Machine learning model to predict milk productivity based on EO data | ||||||||
Product Category | ||||||||
|
| |||||||
Financial Domains | ||||||||
| ||||||||
User requirements | ||||||||
UN9: Understanding stock levels and monitoring supply chains. | ||||||||
Description | ||||||||
Milk and cattle (in weight) productivity have a very strong correlation with multiple factors including pasture quantity and climate data. EO can provide continuous spatial and temporal climatic data such as precipitation, temperature, wind, and evapotranspiration. Also, vegetation indices and biophysical variables derived from satellite data can be used as indicators for pasture quantity. Using machine learning algorithms, EO data can be correlated with historical milk and cattle productivity. Subsequently, these models can be used to predict and estimate productivity using EO data as predictors. | ||||||||
Spatial Coverage Target | ||||||||
Individual farm level | ||||||||
Data Throughput | ||||||||
Rapid tasking Data availability |
|
| ||||||
PRODUCT SPECIFICATIONS | ||||||||
Main processing steps | Climate data such as precipitation, temperature, wind speed and direction, pressure, and humidity can be derived from ERA5-land. Potential evapotranspiration can be calculated from data obtained from ERA5-land. Vegetation indices (such as NDVI, REPO, NDMI, NDCI, and PSRI), biophysical variables (such as LAI), and albedo can be derived from Sentinel-2 or Sentinel-3 based on the application. Green biomass data can be derived from LAI and phenology stages as described previously. By using feature selection algorithms, we can identify and select the most correlated features to milk and cattle productivity to be used as inputs to different machine learning models. After training and validation of different models, we can choose the models with the best performance to estimate and predict milk and cattle productivity based on input EO data. | |||||||
Input data sources | Optical: Sentinel-2&3 Radar: N.A Reanalysis products: ERA5-land Supporting data: Historical milk and cattle (in weight) productivity | |||||||
Accessibility | Sentinel-2&3: freely and publicly available from ESA. ERA5-land: freely and publicly available from EMCWF. | |||||||
Spatial resolution | Sentinel-2: 10 m Sentinel-3: 300 m ERA5-land: 0.1° | |||||||
Frequency (Temporal resolution) | Sentinel-2: 6 days Sentinel-3: Daily ERA5-land: Daily | |||||||
Latency | Daily | |||||||
Geographical scale coverage | Globally | |||||||
Delivery/ output format | Data type: Raster File format: GeoTIFF | |||||||
Accuracies | Thematic accuracy: N.A Spatial accuracy: N.A | |||||||
Constraints and limitations |
| |||||||
User's level of knowledge and skills to extract information and perform further analysis on the EO products. | Skills: Essential Knowledge: Essential | |||||||
Similar Products |
P06: Milk and cattle (in weight) productivity estimation | |
Maturity score | |
Mean: 1.9 | STD: 0.64 |
Constraints and limitations · Cloud presence · Lack of historical milk and cattle (in weight) productivity data to train the models | |
Relevant user needs UN9: Understanding stock levels and monitoring supply chains. | |
R&D gaps · Livestock can be stored in livestock barns and fodder can be imported. The proposed correlation is tricky. · The data exchange between the user and the EO provider. · Data and correlation vary a lot between countries and regions | |
Potential improvements drivers · More forums on this issue · Investigation of the product in different regions | |
Utilisation level review | |
Utilisation score | |
Mean: 2.4 | STD: 0.80 |
No utilisation: Low utilisation Unawareness of the existence of commercial EO products with better specifications Medium utilisation High utilisation | |
Critical gaps related to relevant user needs | |
|
This page has no comments.