📄 medicine

Predicting the distribution of 16 tick species in Japan and its association with the occurrence of tick-borne diseases

By combining nationwide field surveys with species distribution modeling, this study established the first comprehensive maps of 16 tick species across Japan, revealing that snow depth is a key environmental determinant and demonstrating significant correlations between the predicted occupancy rates of specific vector ticks and the occurrence of associated tick-borne diseases.

Original authors: Mebuki Ito, Yuma Ohari, Yuki Ohsugi, Yurie Taya, Naoki Hayashi, Shohei Ogata, Kodai Kusakisako, Yongjin Qiu, Nariaki Nonaka, Keita Mizuma, Shiho Torii, Masahiro Kajihara, Hokkaido University Tick Hunt

Published 2026-07-15

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Mebuki Ito, Yuma Ohari, Yuki Ohsugi, Yurie Taya, Naoki Hayashi, Shohei Ogata, Kodai Kusakisako, Yongjin Qiu, Nariaki Nonaka, Keita Mizuma, Shiho Torii, Masahiro Kajihara, Hokkaido University Tick Hunting Team, Keita Matsuno, Ryo Nakao

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine Japan as a giant, sprawling puzzle made of 47 different pieces (prefectures), stretching from the icy north to the tropical south. For a long time, scientists only had tiny, blurry snapshots of where ticks lived in this puzzle. They knew some ticks were here, and some were there, but they didn't have the full picture. This study is like finally assembling the whole puzzle to see exactly where 16 different types of ticks hang out across the entire country.

The Great Tick Hunt
Between 2013 and 2025, a massive team of researchers (the "Hokkaido University Tick Hunting Team" and friends) went on a nationwide scavenger hunt. They visited 1,018 different spots across 39 prefectures. Using a simple but effective trick—dragging a piece of flannel fabric through the grass to catch the little critters—they collected a whopping 22,416 ticks.

The Crystal Ball: Predicting Where Ticks Live
The team didn't just count the ticks they found; they used a super-smart computer program called Maxent to act like a crystal ball. They fed the program data about the weather, the height of the land, and what the forests looked like. The goal? To predict where ticks could live, even in places the team hadn't visited yet.

The computer learned some fascinating rules about tick real estate:

The Snow Rule: For six of the tick species, the amount of snow in winter was the biggest boss of all. Deep snow actually helps some northern ticks survive (like a warm blanket), but it's a no-go zone for southern ticks.
The Forest Factor: Most ticks love forests, whether they are broadleaf or pine.
The Height Factor: Higher mountains are generally less inviting for many tick species.

The Big Reveal: North vs. South
The maps the team created show a clear split in the tick world:

The Northern Crew: Ticks like Ixodes pavlovskyi and Ixodes persulcatus are the kings of the cool north and high mountains. They are the ones you'd find in Hokkaido and the snowy peaks of Honshu.
The Southern Crew: Ticks like Haemaphysalis flava and Amblyomma testudinarium prefer the warmer, southern regions.

Connecting the Dots: Ticks and Sickness
Here is the most important part: The team wanted to know if these maps could help predict sickness. They compared their tick maps with the official records of five specific tick-borne diseases reported in Japan.

The results were a strong match:

Northern Diseases: The diseases that mostly show up in the north (Lyme disease, Tick-borne encephalitis, and Relapsing fever) lined up perfectly with the maps of the northern Ixodes ticks. The computer suggests that where these ticks are, these diseases are likely to be found.
Southern Diseases: The diseases that hit the south (Severe fever with thrombocytopenia syndrome, or SFTS, and Japanese spotted fever) matched up with the southern ticks like Haemaphysalis flava.

How Sure Are We?
The team didn't just guess; they tested their crystal ball. They checked their predictions against old records of ticks found by other scientists from 1911 to 2021. For 10 out of the 16 tick species, the computer's predictions were very accurate, hitting the "good" thresholds for reliability.

However, the paper is careful to note that for some diseases, like Tick-borne encephalitis (TBE), there were very few reported cases (only 8 in the dataset). Because the data is so sparse, the team suggests their connection for TBE is promising but comes with a lot of uncertainty. They also point out that while the maps show a strong link, they can't explain why a person gets sick just by looking at a map; other factors like how many deer are around or how many people visit the woods also matter.

The Bottom Line
This study didn't just find ticks; it built the first-ever, high-definition map of where 16 different tick species live across all of Japan. By showing that these maps line up with where diseases actually happen, the team suggests these maps are a powerful new tool. They aren't a magic cure, but they are a vital guide for understanding where tick-borne threats might pop up next, helping communities stay one step ahead.

Technical Summary: Predicting the Distribution of 16 Tick Species in Japan and its Association with the Occurrence of Tick-Borne Diseases

Problem Statement
Tick-borne diseases represent a significant public health threat in Japan, characterized by strong spatial heterogeneity due to the country's diverse climatic zones ranging from subarctic to subtropical. While specific diseases like Severe Fever with Thrombocytopenia Syndrome (SFTS) and Japanese Spotted Fever (JSF) are prevalent in warmer southern regions, others such as Lyme Disease (LD), Tick-Borne Encephalitis (TBE), and Relapsing Fever (RF) are largely restricted to cooler northern areas. Despite the critical need for comprehensive distribution data to anticipate disease risk, existing tick survey data in Japan are fragmented, localized, and often employ inconsistent sampling methods. This lack of nationwide data hinders effective public health preparedness. The study addresses the limitation of extrapolating local models to the entire archipelago by aiming to generate the first comprehensive, nationwide distribution models for 16 major tick species and to evaluate their utility in predicting the occurrence of five notifiable tick-borne diseases.

Methodology
The study employed a multi-step approach combining extensive field data collection with Species Distribution Modeling (SDM) and statistical regression analysis:

Data Collection: Ticks were collected using the flagging method at 1,018 sites across 39 prefectures in Japan between 2013 and 2025. A total of 22,416 ticks belonging to 23 species were identified. Presence data were spatially filtered to avoid pseudo-replication, retaining 16 species with at least 10 presence sites for modeling.
Environmental Variables: Twenty-eight environmental variables (climate, topography, and landscape) were initially considered. To mitigate multicollinearity, variables were reduced to 17 using Variance Inflation Factor (VIF) analysis (threshold < 10). Key variables included 19 bioclimatic variables, yearly maximum snow depth, elevation, slope, and land cover percentages (e.g., broadleaf forest, urban cover).
Species Distribution Modeling (SDM): Maxent (Maximum Entropy) models were used to predict potential suitable habitats at a 1 km resolution.
- Bias Correction: A smoothed bias surface based on Gaussian kernel density estimation of sampling sites was used to weight background point selection, accounting for regional variations in sampling effort.
- Model Selection: Models were trained on 75% of the data and tested on 25%. Performance was evaluated using AUC, True Skill Statistic (TSS), Continuous Boyce Index (CBI), and $\Delta$ AUC (difference between training and test AUC) to detect overfitting. Explanatory variables were sequentially removed to optimize model performance based on hierarchical criteria prioritizing low $\Delta$ AUC and high CBI.
Validation: Predicted distributions were validated against an independent dataset of historical tick collection records (1911–2021) compiled by Morishima et al. (2025).
Disease Association Analysis:
- Spatial Overlap: Schoener's D and Warren's I metrics were calculated to assess spatial overlap between tick occupancy rates and disease incidence rates (SFTS, TBE, JSF, LD, RF) at the prefectural level.
- Regression Analysis: Generalized Linear Models (GLM) and Generalized Additive Models (GAM) were fitted to examine the relationship between tick occupancy rates and disease case numbers. Model selection accounted for overdispersion (using Negative Binomial distributions) and spatial autocorrelation (using spatial smooths for SFTS and JSF).

Key Contributions and Results

Nationwide Distribution Maps: The study successfully generated the first comprehensive distribution models for 16 tick species across the Japanese archipelago. The models demonstrated robust performance, with most species achieving AUC > 0.7 and TSS > 0.4.
Environmental Determinants: The analysis identified yearly maximum snow depth as a fundamental determinant of tick distribution, yielding the highest percentage contribution for six species.
- Southern Species: Species with southern distributions (e.g., Haemaphysalis spp.) showed a negative correlation with snow depth.
- Northern Species: Species with northern distributions (e.g., Ixodes spp.) showed a positive correlation with snow depth, likely due to thermal insulation benefits for overwintering.
- Other significant variables included forest cover (positive effect) and elevation (negative effect for many Haemaphysalis species).
External Validation: Ten of the 16 species demonstrated good predictive performance when validated against independent historical records, confirming the robustness of the modeling approach.
Disease Correlation: Regression analyses revealed significant correlations between the occupancy rates of specific vector ticks and the reported cases of five diseases:
- Northern Diseases (LD, RF, TBE): The distributions of Ixodes pavlovskyi and Ixodes persulcatus aligned closely with the occurrence of these diseases. I. pavlovskyi and I. persulcatus were the top predictors for LD and RF, respectively.
- Southern Diseases (SFTS, JSF): Haemaphysalis flava was identified as the primary predictor for JSF, while H. flava and Amblyomma testudinarium were strong predictors for SFTS.
- The regression results for SFTS and JSF were consistent with previous molecular surveillance data where these pathogens had been isolated or detected in the corresponding tick species.

Significance and Claims
The authors claim that this study provides the first comprehensive spatial overview of tick distributions in Japan, filling a critical gap in epidemiological data. The primary significance lies in the demonstration that predicted tick occupancy rates serve as strong predictors for the occurrence of endemic tick-borne diseases.

The study posits that the generated maps serve as basic information for anticipating the occurrence of tick-borne diseases in the region. By establishing the relationship between specific vector distributions and disease incidence, the models offer a practical tool for regional preparedness. The authors note that while the study operates at a prefectural scale and does not incorporate micro-scale factors (such as human activity patterns or vertebrate reservoir distributions), the strong alignment between the macro-scale distribution models and actual disease occurrence validates the utility of these maps for public health planning. Furthermore, the authors suggest that these maps can assist in identifying potential vectors and high-risk areas should emerging tick-borne pathogens be detected in the future.

Technical Summary: Predicting the Distribution of 16 Tick Species in Japan and its Association with the Occurrence of Tick-Borne Diseases

More like this