A landscape composed of people, agriculture, and agroforests

The Western Ghats of India is an important biodiversity hotspot, especially for frogs. It is often described as one of the “hottest” hotspots because of its extraordinary biodiversity and endemism, which coexist with dense human populations. However, only around 10% of the region is under formal state protection, and much of the landscape outside these protected areas has been transformed primarily by open agriculture, agroforests and human settlements. Among these land uses, coffee is one of the most widely cultivated crops, grown extensively across the mid-elevation hills of the Western Ghats. Within this human-modified landscape, coffee agroforests can occupy an exceptionally important role for frog conservation, as coffee is often grown beneath a canopy of native shade trees (1). In landscapes where intact forest has already been cleared or fragmented, studies have shown that well-shaded coffee agroforests can provide structurally complex habitats for mammals, birds, frogs, and reptiles (2).

A three-panel composite labeled robusta coffee, arabica coffee and rainforest: an open canopy with jackfruit hanging from a trunk above coffee bushes, a shadier stand of coffee beneath mid-height trees, and tall closed-canopy forest in misty light
Figure 1. Representative images of the three habitats compared in our study: robusta agroforests with a more open canopy, arabica agroforests with relatively high canopy cover and secondary rainforest with its dense and diverse canopy. Photo: Vijay Karthick

Across the tropics, including the Western Ghats, coffee landscapes are changing rapidly (3). Global coffee cultivation is dominated by two species: arabica (Coffea arabica) and robusta (Coffea canephora). Arabica is increasingly being replaced by robusta, a less shade-dependent crop often associated with more simplified canopies. At the same time, many coffee-growing regions are becoming drier (i.e., experiencing reduced mean annual rainfall). Yet, we still know little about what these changes can mean for frogs. Do arabica and robusta agroforests support the same frog communities? And does their conservation value change between relatively wetter and drier regions of coffee cultivation?

Counting frogs through rainforests and coffee agroforests

Our study took place in the central Western Ghats, across the coffee-growing districts of Hassan and Chikmagalur (4). Much of the rainforest in the study landscape has been replaced by coffee agroforests over the past two centuries. Robusta cultivation has expanded by 37%, while arabica has declined by 6%, driven by a mixture of market conditions, pests, labour, and a changing climate. Both arabica and robusta are also grown across a broad rainfall gradient. We therefore considered arabica and robusta agroforests in both wetter areas, receiving roughly 3,000-4,000 mm of rainfall annually (arabica-wet and robusta-wet), and relatively drier areas receiving 2,000-3,000 mm (arabica-dry and robusta-dry).

A three-part map: an inset locating the Western Ghats within South Asia, a map of peninsular India with the range shaded along the west coast, and a detailed land-cover map showing agroforest, wet evergreen forest and water bodies, dashed rainfall isohyets from 2,000 to 5,000 millimeters, and the coffee and rainforest sampling sites
Figure 2. The study landscape in the central Western Ghats, India. The map on the right shows major land-cover types, rainfall gradient and frog sampling sites. Areas marked as agroforest are mainly shade-coffee agroforests. Source: Adapted from Karthick et al. (2026), Biological Conservation, under CC BY-NC 4.0.

We investigated how different coffee-cropping systems (arabica and robusta) and climatic conditions (relatively wet and dry) shape frog communities in the Western Ghats. During the 2023 monsoon, we surveyed frogs along 43 line transects in one secondary rainforest and six coffee sites, including both arabica and robusta agroforests in wetter and drier parts of the landscape. Altogether, we walked 12.9 kilometres and recorded 2,185 adult frogs representing 28 species. All but two species were endemic to the Western Ghats. In addition to looking at how many frog species (taxonomic diversity) occurred in each habitat, we also examined the variety of functional traits (functional diversity) and evolutionary lineages (phylogenetic diversity) they represented, as well as the occurrence of threatened and sensitive functional groups. This allowed us to ask whether coffee agroforests were conserving the same kinds of frog diversity and species as rainforests, and how reduced rainfall could amplify these changes.

Coffee agroforests in the Western Ghats were full of frogs. We then looked closer.

Coffee agroforests had as many frogs as rainforests. Species richness (number of species in each habitat) did not differ considerably either. However, a closer look revealed a different story. The secondary rainforest supported a distinct frog community and had the highest taxonomic, functional, and phylogenetic diversity. In the rainforest, individuals were more evenly distributed across species, allowing a wider representation of functional types and evolutionary lineages. Coffee communities, by contrast, were often dominated by a smaller group of disturbance-tolerant species such as Clinotarsus curtipes (Bicoloured Frog), Raorchestes luteolus (Blue-eyed Bush Frog), and Pseudophilautus wynaadensis (Wayanad Bush Frog).

A small orange-brown frog with a dark-rimmed eye perched in profile on a broad glossy green leaf against a dark background
Figure 3. Raorchestes luteolus (Blue-eyed Bush Frog), a terrestrial breeding frog found more commonly in the coffee agroforests of central Western Ghats, India. Photo: Vijay Karthick
Two stacked panels: five rank-abundance curves, one per habitat, each labeled with the commonest species codes, and below them an NMDS ordination in which the secondary rainforest ellipse sits well apart from the four overlapping coffee ellipses
Figure 4. Frog abundances and community composition across rainforest, arabica and robusta in wetter and drier sites. (a) The rank-abundance curves show whether abundances were evenly distributed or dominated by a few common species across the habitats, (b) NMDS plot of community composition, illustrating that rainforest supported a distinct frog community compared to coffee agroforests. Source: Reproduced from Karthick et al. (2026), Biological Conservation, under CC BY-NC 4.0.

Arabica, robusta, and a drying landscape

Five panels of point estimates with confidence bars, color-coded by habitat: three panels of taxonomic, functional and phylogenetic diversity in which secondary rainforest sits highest, and two panels of beta-coefficients grouped by breeding strategy and by conservation priority
Figure 5. The secondary rainforest supported the highest taxonomic, functional and phylogenetic diversity, while arabica generally retained more diversity than robusta agroforests (a, b, and c). Occurrences of stream-breeding (lotic) and high conservation-priority frogs were most strongly associated with the secondary rainforest, while pond-breeding (lentic) species were more common in coffee (d and e). Points show average estimates, and bars show 95% confidence intervals. Source: Adapted from Karthick et al. (2026), Biological Conservation, under CC BY-NC 4.0.

The contrast between arabica and robusta agroforests was more subtle than a simple good-coffee, bad-coffee story. Frog abundance was high in both coffee cropping systems, with species richness often comparable. Yet, arabica agroforests generally exhibited greater multidimensional (taxonomic, functional, and phylogenetic) diversity than robusta agroforests. This suggests that while a shift from arabica to robusta agroforests may leave the overall number of frogs largely unchanged, it could reduce the diversity of species, functional traits, and evolutionary lineages represented within the community.

In general, coffee agroforests supported many endemic frogs (but not threatened), including terrestrial breeders considered sensitive to habitat change and moisture availability.

In both coffee systems, drier-zone coffee generally supported lower multidimensional diversity than wetter sites. Stream breeders, terrestrial breeders, and high conservation priority (i.e., threatened and endemic) species also tended to occur less often in drier coffee sites.

Of losers and winners

A small dark brown frog flecked with cream markings and a copper-ringed eye, standing in profile on a wet green leaf at night
Figure 6. Micrixalus elegans (Elegant Dancing Frog), which was detected only in the rainforests. Photo: Vijay Karthick

Stream breeding (lotic) frogs were among the species most affected by the shift from rainforest to coffee. Species belonging to the genera Micrixalus (Dancing Frogs) and Nyctibatrachus (Night Frogs) represent evolutionary lineages roughly 60 million years old and are closely associated with shaded, rocky, flowing streams. In our study, stream breeding frogs were much more likely to occur in the secondary rainforest than in either coffee agroforestry systems. The availability and quality of microhabitat offered a possible explanation: streams and riparian vegetation were numerous and intact in the rainforest, whereas streams in coffee agroforests were often modified for water storage and supply. Rocks and large boulders, which provide important microhabitats for stream-associated frogs, were often absent from streams in coffee agroforests. Other species with strong forest associations included Raorchestes hassanensis (Hassan Bush Frog), Duttaphrynus parietalis (Ridged Toad), and Duttaphrynus brevirostris (Short-nosed Toad). Threatened and range-restricted species (high conservation priority species) also showed stronger associations with the rainforest.

A shallow stream running in a narrow channel over wide exposed pink bedrock in a coffee agroforest, with a person standing on the rock with their back to the camera and a white pickup truck parked among trees beyond
Figure 7. A degraded stream habitat within a coffee agroforest during the monsoon. Photo: Vijay Karthick

By contrast, frogs that reproduce in still water emerged as the apparent “winners”, occurring more frequently in coffee agroforests. Irrigation ponds and temporary roadside pools within coffee agroforests provide abundant breeding sites for them. These artificial ponds can be valuable for conserving stagnant-pool (lentic) breeding species, but they do not replace the fast-flowing, rocky streams required by specialised forest species.

A three-panel composite: a moss-stained concrete water tank holding dark water and leaf litter among coffee bushes, a close-up of a single tadpole on wet algae-covered concrete, and a broad irrigation pond reflecting a pink sunset
Figure 8. Human-made breeding sites within coffee agroforests. Cemented tanks beneath the shade tree canopy collect rainwater and leaf litter during the monsoon (a), creating a suitable habitat for tadpoles, especially for frogs that build foam nests (b). Larger irrigation ponds also serve as breeding sites for several frog species (c). Photo: Vijay Karthick

Protect, restore, and let coffee help forests return

The major highlight of our study is that coffee agroforests matter for frogs - but they cannot substitute for rainforest. Coffee expansion is a widespread driver of tropical deforestation, with extensive forest loss documented across major coffee-producing countries such as Brazil and Vietnam (5).

Remaining primary forests are irreplaceable and must be protected, while secondary rainforests deserve greater attention. The rainforest in our study was once selectively logged, and parts of it were used for shade-coffee cultivation in the past. Following protection, it recovered a structure resembling a lightly disturbed rainforest and now supports stream specialists, threatened endemics and a more diverse frog community. Restoring secondary rainforests and abandoned coffee fields could therefore provide important refuges for frogs.

Climate is another growing concern. Declining rainfall could reduce frog diversity across tropical coffee-growing regions, particularly affecting stream and terrestrial breeding frogs and species of high conservation priority. Previous research in the Western Ghats similarly found lower occupancy among several terrestrial breeding frogs during a dry monsoon year (6).

Management practices within coffee agroforests also matter. Our results show that the ongoing shift from arabica to robusta agroforests could reduce multiple dimensions of frog diversity. Identifying and protecting important frog microhabitats will therefore be essential for sustaining diverse frog communities in coffee-growing landscapes.

On the brighter side, coffee agroforests can act as an important complementary habitat for endemic, terrestrial, and pond-breeding frogs, while also helping restore nearby forests. A recent study in the same landscape led by my supervisor, Anand M Osuri, identified 30 conservation-priority tree species growing in coffee agroforests that could support restoration (7). Instead of being removed during routine weeding, naturally regenerating native seedlings in coffee agroforests can be rescued and utilised to restore degraded forests. This way, coffee agroforests can contribute meaningfully to both biodiversity conservation and rainforest restoration in production landscapes. Such actions will become even more important as coffee cultivation and climate continue to change together.

Dense coffee bushes in the foreground beneath tall slender shade trees, with a lilac and pink twilight sky and distant ridges visible through a gap in the canopy
Figure 9. From a distance, shade-coffee agroforests in the Western Ghats can look almost like a forest. As the sun sets during the monsoon season, coffee agroforests in the Western Ghats reverberate with frog calls. Photo: Vijay Karthick

So, do frogs like coffee as much as we do? Some clearly do, but others are quietly disappearing. The future of frogs in a coffee country depends not only on whether the nights stay alive with calls, but on whose voices make up that chorus.

Figure sources and license

Figures 2 and 5 are adapted, and Figure 4 is reproduced, from the underlying study: Karthick, V., Dass, S., Sadekar, V., Naniwadekar, R., & Osuri, A. M. (2026). Land use and climate shape amphibian multidimensional diversity and conservation in a coffee agroforestry landscape, Western Ghats. Biological Conservation, 322, Article 112057. https://doi.org/10.1016/j.biocon.2026.112057

The version of record is published under the Creative Commons Attribution-NonCommercial 4.0 International license (CC BY-NC 4.0). License terms: https://creativecommons.org/licenses/by-nc/4.0/. Figures 2 and 5 have been modified for this article; Figure 4 is unmodified. This reuse is noncommercial. Attribution here does not imply endorsement of the Amphibian Survival Alliance or of this article by Biological Conservation, its publisher, or the study's co-authors.

Acknowledgements

I am thankful to my supervisors, Anand M. Osuri and Rohit Naniwadekar for their guidance and support throughout the study. The generous support of many coffee agroforest owners and managers, their families, forest department authorities, and our field team from the Nature Conservation Foundation made this study possible. I am also deeply grateful to everyone who helped with logistics, joined us during long nights of fieldwork, and shared their knowledge and ideas. This work was funded by Rohini Nilekani Philanthropies, On the Edge Conservation, Cholamandalam Investment and Finance Company, and Arvind Datar.

References

  1. Sankararaman, V., Dalvi, S., W. Miller, D. A., & Karanth, K. K. (2021). Local and landscape characteristics shape amphibian communities across production landscapes in the Western Ghats. Ecological Solutions and Evidence, 2(4), e12110. https://doi.org/10.1002/2688-8319.12110
  2. Wynter, V., Milner-Gulland, E., & Poore, J. (2025). A global comparison of the biodiversity impacts of coffee agricultural systems–From monoculture to diverse agroforestry. Agricultural Systems, 229, 104449. https://doi.org/10.1016/j.agsy.2025.104449
  3. Harvey, C. A., Pritts, A. A., Zwetsloot, M. J., Jansen, K., Pulleman, M. M., Armbrecht, I., Avelino, J., Barrera, J. F., Bunn, C., Hoyos Garcia, J., Isaza, C., Munoz Ucros, J., Perez-Aleman, C. J., Rahn, E., Robiglio, V., Somarriba, E. J., & Valencia, V. (2021). Transformation of coffee-growing landscapes across Latin America. A review. Agronomy for Sustainable Development, 41, Article 62. https://doi.org/10.1007/s13593-021-00712-0
  4. Karthick, V., Dass, S., Sadekar, V., Naniwadekar, R., & Osuri, A. M. (2026). Land use and climate shape amphibian multidimensional diversity and conservation in a coffee agroforestry landscape, Western Ghats. Biological Conservation, 322, Article 112057. https://doi.org/10.1016/j.biocon.2026.112057
  5. Robusta’s Reckoning. (2026, July 23). Coffee Watch. https://coffeewatch.org/vietnams-robustas-reckoning/
  6. Sankararaman, V., & W. Miller, D. A. (2024). Life-history traits govern the habitat use of diverse amphibian assemblages in an agroforest landscape matrix. Animal Conservation, 27(1), 86-97. https://doi.org/10.1111/acv.12882
  7. Osuri, A. M., Kumar, V. S., Dutta, V., Vidwath, S. M., Cariappa, K. M., Purnesh, C. D., & Shetty, S. (2025). Shade coffee agroforests as native plant sources for restoration. Journal of Applied Ecology, 62(6), 1330-1336. https://doi.org/10.1111/1365-2664.70046