Vaisala Oyj Class A · OMXHEX:VAIAS
Vaisala Oyj Class A — Investor Presentation (May 2023)
36 slides from Vaisala Oyj Class A's investor presentation, May 2023
Slide 1: Vaisala Xweather Global Air Quality Index
Vaisala Xweather Global Air Quality Index Air quality index and pollutant components with 72h forecast 1
Slide 2: Weather and environment expertise since 1936
Weather and environment expertise since 1936 Net sales in 2022 12% 2,200+ 25+ 514.2 of net sales employees offices in MEUR invested into R&D worldwide 17 countries 300+ Business in over Employee engagement index 100% product renewable families 150 4.2 / 5 electricity countries 2
Slide 3: 9 out of 10 people worldwide breathe polluted air
9 out of 10 people worldwide breath polluted air 4th most common reason for death after high blood pressure, smoking and nutrition Ambient air pollution contributes to 5.4% of all deaths worldwide 4.2 million deaths every year 3
Slide 4: From reactive to proactive
From reactive to proactive Monitoring Mitigation Warnings Smart city With better information, people become more aware, active, and committed. 4
Slide 5: AQI Forecast is based on various elements
AQI Forecast is based on various elements In-situ measurements Remote observations to Physical modelling for AI, ML and Intelligent air quality for local observations understand the condition initial data fusion and advanced data and alerting to of the atmosphere forecasting algorithms make decision
Slide 6: Global forecast with unprecedented accuracy
Global forecast with unprecedented accuracy 72-hour forecast in 1-hour intervals Dynamic spatial resolution from 2–10 km, typically 5 km Learns from actual measurement data, calibrates itself by machine learning methods Combines long range transportation modelling and modern AI methods 6
Slide 7: Continental forecast with outstandingly short latency
Continental forecast with outstandingly short latency Improved forecast accuracy, due to the latency decrease of the observations → traditional models have up to 10 hours of latency 7
Slide 8: Xweather Global Air Quality Index
• Air Quality Index is a Xweather Global Air Quality Index combination of the contributions of several pollutants into single reference number • Vaisala’s Air Quality Index combines superior air quality modeling and local quality observations into global forecast wherever you are • It leverages real
Slide 9: Air Quality Index and individual pollutants
Air Quality Index and individual pollutants Air Quality Index, a composite measure of Index the contributions of individual pollutants Particles Gases CO Carbon monoxide Fine inhalable particles with PM2.5 Ground-level ozone O3 diameters of 2.5 μm or less NO Nitrogen monoxide Inhalable particles wit
Slide 10: Global level of individual pollutants
Global level of individual pollutants PM10 Inhalable particles with diameters of 10 μm or less PM2.5 Fine inhalable particles with diameters of 2.5 μm or less 10
Slide 11: Global level of individual pollutants
Global level of individual pollutants NO2 - Nitrogen dioxide CO - Carbon monoxide 11
Slide 12: Pollutant concentrations and forecast
Pollutant concentrations and • Xweather Global Air Quality forecast Index can provide also all pollutant components (PM2.5, Europe PM10 concentration on 2023-03-15T10_00 PM10, NO2, O3, SO2 & CO) and their 72h forecasts Europe AQI on 2023-03-15T10_00 Example – Dust from Sahara PM10 • Image shows comp
Slide 13: Global AQI – key characteristics
Global AQI – key characteristics • 72 hour forecast in 1-hour steps • Dynamic spatial resolution from 2 to 10 km • Learns from actual measurement data, calibrates itself by machine learning methods • Combines long range transportation modelling and modern AI methods 13
Slide 14: Global air quality (AQI) data flow
Global air quality (AQI) data flow Data Collection Module Observations Long Range Weather data GIS Data Data transport models models Data Transformation layer Model execution Data delivery layer 14
Slide 15: Global air quality (AQI) data collection module
Global air quality (AQI) data collection module Data Collection Module Observations Long Range Weather data GIS Data Data transport models models China Direct India Direct Per country reference stations More than 10.000 stations globally Customer own Vaisala AQT stations 15
Slide 16: Local information improves the global forecast
Local information improves the • Scale up spatial resolution global forecast achieving from 2 to 5km Vaisala stations • Improved forecast accuracy, due to the latency decrease of the observations → traditional models have up to 10h of latency • Observations enable to provide hourly updates → traditi
Slide 17: Global air quality (AQI) data collection module
Global air quality (AQI) data collection module Data Collection Module Observations Long Range Weather data GIS Data Data transport models models • Over 10.000 observations taken in now and growing • Global coverage extension program, with continuous update of observations. • Continuous monitoring o
Slide 18: Global air quality (AQI) data collection module
Global air quality (AQI) data collection module Data Collection Module Observations Long Range Weather data GIS Data Data transport models models SILAM CAMS ENSEMBLE GENAQ (PL) EURAD-IM (DE) CHIMERE (FR) MOCAGE (FR) MONARCH (SP) DEHM (DK) MINNI (IT) MATCH (SE) EMEP (NO) LOTOS-EUROS (NL) SILAM (FI) C
Slide 19: Global air quality (AQI) data collection module
Global air quality (AQI) data collection module Data Collection Module Observations Long Range Weather data GIS Data Data transport models models **To be added Global forecast Model (GFS) Vaisala Weather API 19
Slide 20: Global air quality (AQI) data collection module
Global air quality (AQI) data collection module Data Collection Module Observations Long Range Weather data GIS Data** Data transport models models **To be added Satellite Land Elevation Population Road traffic Building height Land use 20
Slide 21: Global air quality (AQI) data flow
Global air quality (AQI) data flow Data Collection Module Observations Long Range Weather data GIS Data Data transport models models Data Transformation layer Model execution Data delivery layer 21
Slide 22: Integrated AI/ML with end-to-end air quality modelling
Integrated AI/ML with end-to-end air quality modelling • AI/ML-native implementation as an integral part of the Xweather Global Air Quality solution • End-to-end approach from observations to AI/ML and up to forecast • Best spatiotemporal resolution and frequency of updates, while ingesting multi-so
Slide 23: AI/ML Ops pipeline
AI/ML Ops pipeline MLOPS BASE line • MLOPS is AI/ML-native and integral part of the Xweather global air quality solution • The implementation leverages AWS capabilities • S3 (storage), CloudWatch (logs), Elastic Container Service (ECS) for computations, Glue (data integration), Lambda for less- dema
Slide 24: Quality control, validation, and continuous improvement
Quality control, validation, and continuous improvement • Model performance is monitored constantly against observations. • Overall results of the analytics is reported to customers. • Advanced performance analytics is used for continuously improving the model. • AQI has 2 years roadmap for further
Slide 25: Validation: How much better are we? Concentrations
Validation: How much better are we? Concentrations 25
Slide 26: Validation: How much better are we? Air quality Index
Validation: How much better are we? Air quality Index 26
Slide 27: Global Air Quality
Global Air Quality 27
Slide 28: Xweather Global Air Quality
Xweather Global Air Quality World CO [g/kg] _ 20230210_12:00 • Xweather Global Air Quality Index is currently using more than 10,000 observations stations globally • PM2.5, PM10, NO, NO2, O3 & CO plus Index and forecast for all Global station coverage World PM 2.5 [g/m3] 20230209_22:00 28
Slide 29: Xweather Global Air Quality
Xweather Global Air Quality Regional observation coverage • Enhanced regional air quality index United States Regional station coverage using information from key locations • We are expanding the observations used globally and we can prioritize key locations for you Germany United Kingdom • In the p
Slide 30: Per-region AQI calculations
• Different AQI calculation Per-region AQI calculations methods currently available and possibility to add more Method Name AQI method methods upon request USA EPA AirNow USAEPA_AirNow Maximum European AQI (EAQI) EEA_EAQI Average European Common AQI (CAQI) EEA_CAQI Average • Based on pollutant Germa
Slide 31: Region AQI calculations
Region AQI calculations USA AirNow Germany UBA Daily AQI 0-50 Good 0-1 Very good 51-100 Moderate 2-3 Good Unhealthy for 101-150 sensitive groups 3-4 Moderate 151-200 Unhealthy 4-5 Poor 201-300 Very unhealty 5-6 Very poor >300 Hazardous European AQI UK DAQI 0-25 Good 1-3 Low 25-50 Fair 4-6 Moderate 5
Slide 32: Roadmap
Roadmap Continuous improvements bringing the most value to your state 34
Slide 33: AQI roadmap
Global AQI roadmap Global observation GIS data integration integration Lightning Nowcasting Model Integration of North America integration LRTM CMAQ Model Weather model integration Anomaly detection Advanced analytics Dynamic emission inventory Features Customizable AQI level threshold for notificat
Slide 34: Why Vaisala Xweather Air Quality
Why Vaisala Xweather Air Quality Data fusion of huge number of different data sources Based on latest Most advanced advances in Air Quality numerical Index solution weather modeling Aligned with Air Quality Directive 2008/50/EC Culmination of Advanced over 20 years handling of reseach in air emissio
Slide 35: Summary
Summary • Global air quality index wherever you are • Index is calculated based on local standards • Air quality data with 72-hour forecast • Rely on comprehensive, validated forecast • Constant improvements (2 years roadmap) • Aligned with United Nations’ Sustainability Development Goals and the Eu
Slide 36: Thank you
Thank you 38