Throughout my career, Iâve worked at the intersection of climate, geospatial data, and applied machine learning. I build reusable decision tools from climate and geophysical data, with a bias toward workflows that can be reviewed, reproduced, and shipped.
Research credibility
Peer-reviewed work on snow distribution from satellite laser altimetry (2025). See publications.
Climate relevance
Bias correction and spatial downscaling for climate and energy datasets, with uncertainty-aware evaluation.
Shipping ability
Real-time estimation prototype: Unscented Kalman Filter + FastAPI + PostgreSQL, designed as a blueprint for B2B decision tooling.
Estimating the variability of snow depth in remote areas poses significant challenges due to limited spatial and temporal data availability. This study uses snow depth measurements from the ICESat-2 satellite laser altimeter, which are sparse in both space and time, and incorporates them with climate reanalysis data into a downscaling-calibration scheme to produce monthly gridded snow depth maps at microscale (10 m). Snow surface elevation measurements from ICESat-2 along profiles are compared to a digital elevation model to determine snow depth at each point. To efficiently turn sparse measurements into snow depth maps, a regression model is fitted to establish a relationship between the retrieved snow depth and the corresponding ERA5 Land snow depth. This relationship, referred to as subgrid variability, is then applied to downscale the monthly ERA5 Land snow depth data. The method can provide timeseries of monthly snow depth maps for the entire ERA5 time range (since 1950). We observe that the generic output should be calibrated by a small number of localized control points from a one-time field survey to reproduce the full snow depth patterns. Results show that snow depth prediction achieved a R2 model fit value of 0.81 (post-calibration) at an intermediate scale (100 m Ă 500 m) using datasets from airborne laser scanning (ALS) in the Hardangervidda region of southern Norway, with still good results at microscale (R2 0.34, RMSE 1.28 m, post-calibration). Bias is greatest for extremes, with very high/low snow depths being under- and overestimated, respectively. Modeled snow depth time series at the site level have a slightly smaller RMSE than ERA5 Land data, but are still consistently biased compared to measurements from meteorological stations. Despite such localized bias and a tendency towards average snow depths the model reproduces the relative snow distribution pattern very accurately, both for peak snow (Spearmanâs Ï 0.77) and patchy snow meltout in late spring (Matthews correlation coefficient 0.35). The method relies on globally available data and is applicable to other snow regions above the treeline. Though requiring area-specific calibration, our approach has the potential to provide snow depth maps in areas where no such data exist and can be used to extrapolate existing snow surveys in time and over larger areas. With this, it can offer valuable input data for hydrological, ecological or permafrost modeling tasks.
