The research, published in the journal Parasites & Vectors, was conducted by James C. Burtis, Erik Foster, and Rebecca J. Eisen of the CDC’s Division of Vector-Borne Diseases in Fort Collins, Colorado. Their central question was deceptively simple: across seven western states, which counties have environmental conditions suitable both for the bacterium to maintain itself in local wildlife cycles and for infected nymphs to climb up onto vegetation and wait for a host? That second condition matters enormously for human health. Nymphal ticks are tiny, roughly the size of a poppy seed, and their host-seeking behavior above the ground is what brings them into contact with people hiking, gardening, or working outdoors. A tick that never quests above vegetation poses far less risk of a bite.

To build their analysis, the team confronted the fundamental scarcity of western tick surveillance data. Systematic records documenting the densities of host-seeking, infected nymphs are limited in the region, in stark contrast to the eastern United States where such data underpin county-level risk maps. The researchers turned to the ArboNET Tick Module, a national database maintained by the CDC that aggregates tick surveillance submissions from public health partners. Because those records alone could not cover the full geographic extent of potential tick habitat, they supplemented the database with an exhaustive review of the published literature and public health records, collating county-level observations of host-seeking nymphs and separately compiling reports of B. burgdorferi sensu stricto infections detected in both nymphal and adult I. pacificus.

The compilation revealed a striking imbalance. Host-seeking I. pacificus nymphs had been recorded in 54 counties across the seven western states, while evidence of Bbss-infected I. pacificus existed in only 36 counties, all of them in California, Oregon, and Washington. That distinction is critical. A tick population can thrive in a county without the bacterium circulating among the wildlife that feed the ticks. Enzootic transmission, the sustained cycling of the spirochete between ticks and reservoir hosts such as small mammals and birds, requires a specific confluence of vector abundance, competent reservoir species, and favorable climate. Some counties may host abundant ticks but, for reasons of host community composition or microclimate, never sustain the pathogen.

With these records in hand, the researchers built separate habitat suitability models for the two phenomena: one predicting where host-seeking nymphs are likely to occur, and another predicting where Bbss-infected ticks of either stage have been or are likely to be found. The modeling approach drew on multiple algorithms rather than a single statistical technique, a strategy that guards against the idiosyncrasies of any one method. Crucially, the team set the suitability thresholds of all algorithms to achieve 90 percent sensitivity, meaning the models were deliberately tuned to flag nearly every county with a known record, even at the cost of including some counties where the tick or pathogen may ultimately be absent. In public health terms, this conservative choice favors caution: it is better to overestimate the map of potential risk than to tell residents of an affected county they are safe.

The final step was an overlay. By intersecting the suitable habitat for host-seeking nymphs with the suitable habitat for infected I. pacificus, the researchers identified counties where the environment plausibly supports the full combination of conditions that generate acarological risk: infected nymphs actively questing above vegetation. The resulting map told a geographically focused story. Suitable habitat for encountering Bbss-infected host-seeking nymphs clustered mostly in central and northern California and along the coastal corridors of Washington and Oregon. The pattern was markedly narrower than the overall footprint of the tick itself, confirming that I. pacificus occupies a far broader range than the Lyme disease transmission zones within it.

The numbers quantify that contrast vividly. An earlier modeling effort had classified 113 counties across the western states as suitable for I. pacificus. Yet the new analysis found only 76 counties characterized as suitable for detecting Bbss-infected host-seeking nymphs by at least one of the team’s models. In other words, more than a third of the counties that could host the tick lack evidence of conditions supporting the infected, questing nymphs that pose the greatest hazard to people. For residents and clinicians, this reframes the geography of concern: living in a county where western blacklegged ticks exist is not equivalent to living in a county where Lyme transmission cycles are plausible.

The study’s authors are careful about what their models can and cannot say. The suitability maps identify where infected, host-seeking nymphs are environmentally plausible, but they do not differentiate the magnitude of risk within or among counties. A county flagged as suitable might support low densities of infected nymphs while another supports densities orders of magnitude higher, and the binary classification cannot capture that gradient. Nor can the models replace direct measurement; density of infected nymphs, the metric most tightly linked to human case risk, requires labor-intensive field dragging and pathogen testing that surveillance programs in the West have struggled to sustain. What the models provide instead is a prioritization tool, a way to direct scarce surveillance resources toward the counties where field data would be most informative.

That prioritization is precisely how the authors frame the study’s practical payoff. As a next step toward improving acarological risk assessment in the western United States, the research highlights specific counties where enhanced tick surveillance is warranted. Field confirmation in modeled-suitable counties without existing records could verify the presence of infected questing nymphs, while surveillance in borderline counties could sharpen the thresholds of future models. The approach mirrors strategies used in the East, where layered surveillance and modeling have produced the county maps that clinicians and the public routinely consult. Bringing the West closer to that standard has been a long-standing goal of CDC vector-borne disease programs, and the intramurally funded study represents a substantial advance toward it.

For the public, the message embedded in the maps is one of informed vigilance rather than alarm. The core risk zones in central and northern coastal California and the Pacific Northwest coasts are places where outdoor recreation and residential development interleave with tick habitat, and where standard precautions remain essential: using repellents, performing tick checks after time outdoors, and promptly removing attached ticks. At the same time, the finding that infected questing nymphs are plausible in fewer counties than the tick’s total range offers reassurance to residents of inland and southern areas where the tick occurs but the pathogen apparently does not. As climate and land use continue to shift, the interplay between tick distributions and pathogen transmission will evolve, and models like these, anchored to surveillance and repeatedly updated, will be central to keeping western Lyme disease risk maps honest, current, and actionable.

Subject of Research: Habitat suitability modeling for Borrelia burgdorferi-infected western blacklegged tick nymphs in the western United States

Article Title: Identifying counties with suitable habitat for host-seeking Borrelia burgdorferi sensu stricto-infected Ixodes pacificus nymphs across seven states in the western United States

Article References: Burtis, J. C., Foster, E., & Eisen, R. J. (2026). Identifying counties with suitable habitat for host-seeking Borrelia burgdorferi sensu stricto-infected Ixodes pacificus nymphs across seven states in the western United States. Parasites & Vectors. https://doi.org/10.1186/s13071-026-07730-5

Image Credits: AI Generated

DOI: 10.1186/s13071-026-07730-5

Keywords: Lyme disease, Ixodes pacificus, Borrelia burgdorferi, tick surveillance, habitat suitability modeling, nymphal ticks, western United States, CDC, acarological risk, vector-borne disease, county-level mapping, Parasites & Vectors

The study, led by Farhad Sedaghati of Aon Impact Forecasting together with Shahram Pezeshk and Mehran Davatgari-Tafreshi of the University of Memphis, focuses on a quantity called the Effective Amplitude Spectrum, or EAS. Unlike the response spectra most commonly used in building design, which describe how an idealized oscillator reacts to shaking, the EAS is derived directly from the Fourier amplitude spectrum of the recorded motion. That gives it a direct physical interpretation: it reflects how energy is distributed across frequencies by the earthquake source, modified by the path the waves travel and by the soil conditions beneath the recording station. Engineers can also feed EAS predictions into random vibration theory to estimate response spectra quickly, without running time-consuming simulations.

To train their model, the researchers drew on the NGA-West3 database, a large and uniformly processed collection of three-component accelerograms from shallow crustal earthquakes in active tectonic regions. The raw compilation contained 175,284 records, but the team applied a stringent sequence of quality-control filters before a single neuron was trained. Records were excluded if the earthquake was smaller than magnitude 3.0, if the hypocenter was deeper than 30 kilometers, if the rupture distance exceeded 300 kilometers, or if metadata such as fault dip or shear-wave velocity were missing or implausible. Sensors buried more than two meters below the ground surface were rejected, as were stations classified as non-free-field. After all screening steps, 50,639 high-quality records remained, roughly 90 percent of them from California, with the rest contributed by Alaska, Idaho, Nevada, Utah, Washington, and Wyoming.

The neural network itself is deliberately compact. It takes six explanatory variables as input: moment magnitude, rupture distance, fault dip, depth to the top of the rupture, the time-averaged shear-wave velocity in the upper 30 meters of the site, and the depth to a stiff 2.5 kilometers-per-second shear-wave velocity horizon that characterizes sedimentary basins. From these six numbers, the network predicts the natural logarithm of EAS at 22 discrete frequencies spanning 0.05 to about 33 hertz, all at once. This multi-output design proved crucial. When the researchers initially tried predicting each frequency with a separate single-output model, the resulting spectra were jagged and unstable. Predicting all frequencies simultaneously through shared hidden layers allowed the network to learn cross-frequency patterns directly from the data, producing smooth, physically plausible spectra.

One of the thorniest practical problems was missing data. A recording may contain reliable spectral information at some frequencies but not others, because signal-to-noise ratios degrade at both the lowest and highest frequencies. Rather than discarding every record with a single gap, the team used a masked loss function during training. Missing values, represented as not-a-number entries, simply do not contribute to the error that the network tries to minimize, so partially observed spectra still teach the model whatever they can. This masked approach substantially increased the effective amount of data available, particularly at the frequency extremes where coverage is thinnest.

Avoiding inflated estimates of accuracy required equally careful design of the evaluation strategy. Recordings from the same earthquake share source characteristics, so splitting the data randomly would let the network effectively memorize events it was supposed to predict. The researchers therefore partitioned the dataset at the event level, keeping all recordings of a given earthquake together in the same fold. Hyperparameters were selected through ten-fold event-grouped cross-validation, and roughly five percent of events were held out entirely as an independent evaluation set, untouched during architecture selection, preprocessing, or training decisions. After that out-of-event assessment was complete, the final network was retrained on the full retained dataset so that every available earthquake contributed to the distributed model. The winning architecture turned out to be modest: two hidden layers with 32 and 16 neurons, a batch size of 64, and a learning rate of ten to the minus two.

The predictions that emerged show textbook seismological behavior, despite the fact that no functional form was ever imposed. Spectral amplitudes rise monotonically with magnitude and fall with distance, with high frequencies decaying faster than low ones as waves scatter and lose energy in the crust. At low frequencies the spectra steepen toward a corner frequency in a manner consistent with the classical omega-square source model, and at high frequencies they roll off in the way associated with near-surface damping. The team is careful to note that these are qualitative consistencies, not quantitative estimates of source parameters such as corner frequency or kappa, and that predictions above magnitude 7.36, the largest event in the training data, are extrapolations presented for illustration rather than validated forecasts.

Beyond median predictions, the researchers quantified aleatory variability, the irreducible randomness in ground motion that remains after source, path, and site effects are accounted for. Using a sequential mixed-effects decomposition, they separated total residuals into between-event, site-to-site, and event-site-corrected components at each of the 22 frequencies. The standard deviations show clear frequency dependence, increasing toward higher frequencies where ground motion is most sensitive to record-specific details, and they fall within the range reported by existing empirical EAS models. This decomposition matters for practice: in conventional probabilistic seismic hazard analysis the total variability is used directly, but for site-specific studies the explicit separation allows engineers to condition on measured site behavior and reduce uncertainty.

How does the machine stack up against the human-crafted equations it joins? The team compared their model against three established empirical EAS ground-motion models, Bora and colleagues from 2019, Bayless and Abrahamson from 2019, and Campbell and Bozorgnia from 2025, all built on the earlier NGA-West2 database. Across representative scenarios, the neural network’s median predictions track the empirical models closely in both spectral shape and scaling trends, and its variability estimates sit within the published range. The authors are candid that these comparisons demonstrate physical plausibility rather than superiority, since a rigorous head-to-head test would require fitting a parametric model to identical data with identical partitions, something they flag as important future work. Interestingly, the network also reproduced the observed cross-frequency correlation structure of the data with a Pearson correlation of 0.964 between observed and predicted correlation coefficients.

The researchers have released the trained model, its preprocessing scaler, and example code on GitHub, making the tool immediately usable by hazard analysts. Because the network produces broadband Fourier spectra in a fraction of a second, it is well suited to applications that demand thousands or millions of repeated ground-motion evaluations, such as probabilistic seismic hazard analysis and simulation-based risk studies. The model is strictly valid within the range of its training data, magnitudes from about 3.0 to 7.36, distances up to 300 kilometers, and sites from soft soil to hard rock, and it is calibrated specifically for western United States tectonics. Future directions include explainable artificial intelligence techniques to reveal which predictors matter most at each frequency, and physics-informed training that would embed source models such as Brune’s spectrum directly into the learning objective, potentially improving extrapolation to the largest, rarest earthquakes that engineers care about most.

Subject of Research: A data-driven neural network ground-motion model for Fourier amplitude spectra of shallow crustal earthquakes using the NGA-West3 database

Article Title: A data-driven ground-motion model for Fourier amplitude spectra of shallow crustal earthquakes using the NGA-West3 database

Article References: Sedaghati, F., Pezeshk, S., & Davatgari-Tafreshi, M. (2026). A data-driven ground-motion model for Fourier amplitude spectra of shallow crustal earthquakes using the NGA-West3 database. Bulletin of Earthquake Engineering. https://doi.org/10.1007/s10518-026-02708-1

Image Credits: AI Generated

DOI: 10.1007/s10518-026-02708-1

Keywords: ground-motion model, effective amplitude spectrum, NGA-West3, artificial neural network, machine learning, Fourier amplitude spectrum, seismic hazard analysis, shallow crustal earthquakes, aleatory variability, random vibration theory, western United States, earthquake engineering