Harmonic Baseline Experiments

This dataset was developed as part of the following study of harmonic baseline model parameterization for forest condition monitoring:

Pasquarella, V.J., Mickley, J.G., Barker Plotkin, A., MacLean, R. G., Anderson, R. M., Brown, L. M., Wagner, D. L., Singer, M. S., & Bagchi, R. (2021). Predicting defoliator abundance and defoliation measurements using Landsat-based condition scores. Remote Sensing in Ecology and Conservation. https://doi.org/10.1002/rse2.211

We implemented a previously published harmonic modeling approach for forest condition monitoring in Google Earth Engine and systematically assessed the relative ability of condition change products generated using various model parameterizations for predicting pest abundances and defoliation during the 2016-2018 gypsy moth (Lymantria dispar) outbreak in southern New England. We ran a series of 32 experiments that considered a variety of parameter choices for establishing multi-year “baseline” models representing relatively stable forest conditions for each Landsat pixel in our study area. We tested a full set of factors including (a) spectral vegetation index used for model fitting, (b) baseline-modeling period, (c) frequencies of harmonic regression terms, and (d) differences in Landsat time series input imagery. We generated average condition score estimates for each of these 32 baseline parameterizations for a May 1 to September 30, 2017 monitoring period, then used Generalized Linear Mixed Models to test the relationships between ground-based observations of defoliation and defoliator abundance (larva and egg masses).

This archived dataset includes the full set of experimental raster results, as well as a “reanalysis” product from a previous implementation of our condition monitoring workflow. We tested the impact of changing the following parameters on condition score estimates for the Landsat scenes intersecting Southern New England (Massachusetts, Rhode Island, and Connecticut):

Spectral transforms:

TCG = Tasseled Cap Greenness 

NDVI = Normalized Difference Vegetation Index

SR = Simple Ratio

EVI = Enhanced Vegetation Index 

Baseline periods:

2000-2010

2005-2015

Harmonic frequencies:

h12 = 12-month and 6-month harmonics, i.e. 1/365.25, 2/365.25 

h13 = 12-month and 4-month harmonics, i.e. 1/365.25, 3/365.25 

Time series image inputs:

full = All available observations

16d = Single-sensor, excluding Landsat 7 when possible

Parameters used for each experiment are indicated in raster file names, and each results raster includes two bands: (1) score_weighted_mean, which provides an average condition score for the monitoring period weighted across Landsat Paths, and (2) monitor_nobs, which gives the total number of observations used to compute the score_weighted_mean.

Condition scores are calculated based on the difference between observed and predicted "greenness" in a given spectral vegetation index, divided by the root mean squared errors of the baseline model, and thus provide a normalized metric of anomalies in reflectance properties indicative of changes in vegetation condition.

All experiments were generated for the year 2017 with a monitoring period from May 1 to September 30, and we have also included the 2017 assessment produced as part of our "reanalysis" study for comparison. 

An Earth Engine app for viewing these results interactively is available here: https://valeriepasquarella.users.earthengine.app/view/harmonic-baseline-experiments

Additional information on model parameterization rankings can be found in the associated publication (Pasquarella et al. 2021).

Version 1.1 update (2021/05/28) includes separate zipped archives for top model results for defoliation, larval, and egg mass reference datasets (see Pasquarella et al. 2021 for additional information on these datasets). 

Note: In order to provide the greatest level of flexibility in working with condition score data, original experiment results are provided for the full study area with no masking applied. However, condition scores are best suited for assessing vegetation condition changes for forest ecosystems and our assessments are most valid for forested areas. Therefore, recognizing that basic masking may further facilitate use by external users, we now also provide the following versions of the top model datasets within each zipped archive: raw (full results), masked-tcc (pixels with less than 75% canopy cover in the NLCD 2016 Tree Canopy Cover dataset masked), and masked-gsw (pixels with greater than 50% water occurrence based on the v1.2 Global Surface Water Mapping Layers dataset masked). Additionally, we provide results for 2016 and 2018 as well as 2017 for top model parameterizations to enable further comparisons across years. Masked values are recorded as 'nodata' (though alternate versions where masked values are recorded as -9999 are available upon request).

Primary contact: Dr. Valerie Pasquarella ([e-mail])