REINWARDTIA 2017 16 (1) ISSN 0034 – 365 X | E-ISSN 2337 − 8824 | Accredited 792/AU3/P2MI-LIPI/04/2016 A JOURNAL ON TAXONOMIC BOTANY, PLANT SOCIOLOGY AND ECOLOGY REINWARDTIA A JOURNAL ON TAXONOMIC BOTANY, PLANT SOCIOLOGY AND ECOLOGY Vol. 16 (1): 1 – 48, June 15, 2017 Chief Editor Kartini Kramadibrata (Mycologist, Herbarium Bogoriense, Indonesia) Editors Dedy Darnaedi (Taxonomist, Herbarium Bogoriense, Indonesia) Tukirin Partomihardjo (Ecologist, Herbarium Bogoriense, Indonesia) Joeni Setijo Rahajoe (Ecologist, Herbarium Bogoriense, Indonesia) Marlina Ardiyani (Taxonomist, Herbarium Bogoriense, Indonesia) Himmah Rustiami (Taxonomist, Herbarium Bogoriense, Indonesia) Lulut Dwi Sulistyaningsih (Taxonomist, Herbarium Bogoriense, Indonesia) Topik Hidayat (Taxonomist, Indonesia University of Education, Indonesia) Eizi Suzuki (Ecologist, Kagoshima University, Japan) Jun Wen (Taxonomist, Smithsonian Natural History Museum, USA) Barry J Conn (Taxonomist, School of Life and Environmental Sciences, The University of Sydney, Australia) David G. Frodin (Taxonomist, Royal Botanic Gardens, Kew, United Kingdom) Graham Eagleton (Wagstaffe, NSW, Australia) Secretary Rina Munazar Layout Liana Astuti Illustrators Subari Wahyudi Santoso Anne Kusumawaty Correspondence on editorial matters and subscriptions for Reinwardtia should be addressed to: HERBARIUM BOGORIENSE, BOTANY DIVISION, RESEARCH CENTER FOR BIOLOGY– INDONESIAN INSTITUTE OF SCIENCES CIBINONG SCIENCE CENTER, JLN. RAYA JAKARTA – BOGOR KM 46, CIBINONG 16911, P.O. Box 25 CIBINONG INDONESIA PHONE (+62) 21 8765066; Fax (+62) 21 8765062 E-MAIL: reinwardtia@mail.lipi.go.id http://e-journal.biologi.lipi.go.id/index.php/reinwardtia Cover images: Catanthera keris Veldk. (1. Inflorescences; 2. Close up flower; 3. Flower bud), Medinilla squillula Veldk. (4. Habit; 5. Branches; 6. Fascicle of uniflorous Infructescences), Medinilla uninervis Veldk. (7. Habit. Note 1-nerved leaves; 8. infructescence; 9. Immature and mature fruits), Medinilla zoster Veldk. (10. Habit; 11. Inflorescences; 12. Flower). Photo credits: Bangun 223, Lowry & Phillipson 7287, Mahroji, Fabanyo & Soleman 69, Callmander, et al. 1067. 1 2 3 4 5 6 7 8 9 10 11 12 The Editors would like to thank all reviewers of volume 16(1): Agus Susatya - University of Bengkulu, Bengkulu, Indonesia Agus Sutanto - Indonesian Tropical Fruit Research Institute (ITFRI), West Sumatra, Indonesia Axel D. Poulsen - Royal Botanic Garden Edinburgh, Edinburgh, Scotland, UK Andrew Powling - School of Biological Sciences, University of Portsmouth, Portsmouth, UK Elham Sumarga - School of Life Sciences & Technology, Institut Teknologi Bandung, Bandung, Indonesia Meekiong Kallu - University Malaysia Sarawak, Samarahan, Sarawak, Malaysia Harry Wiriadinata - Herbarium Bogoriense, Indonesian Institute of Sciences, Bogor, Indonesia Ulrich Meve - Lehrstuhl für Pflanzensystematik, Universität Bayreuth, Bayreuth, Germany Mien A. Rifai - Akademi Ilmu Pengetahuan Indonesia (AIPI), Jakarta, Indonesia REINWARDTIA Vol 16 No 1, pp: 11 – 18 11 PREDICTING HABITAT DISTRIBUTION OF ENDEMIC AND CRITICALLY ENDANGERED DIPTEROCARPUS LITTORALIS IN NUSAKAMBANGAN, INDONESIA Received July 27, 2016; accepted January 31, 2017 IYAN ROBIANSYAH Bogor Botanic Gardens, Center for Plant Conservation � LIPI, Jl. Ir. H. Juanda 13, Bogor 16112, Indonesia. Email: iyan.robiansyah@lipi.go.id ABSTRACT ROBIANSYAH, I. 2017. Predicting habitat distribution of endemic and critically endangered Dipterocarpus littoralis in Nusakambangan, Indonesia. Reinwardtia 16(1): 11 – 18. —The tree species Dipterocarpus littoralis (Blume) Kurz. is endemic to Nusakambangan and categorized as critically endangered. In the present study, the habitat suitability of the species in Nusakambangan was predicted using logistic regression analysis and Maxent model. Three topographic variables (elevation, slope, and aspect), distance from river and coastline, and one vegetation index (Normalized Dif- ference Vegetation Index (NDVI)) as well as two water content indexes (Normalized Difference Water Index (NDWI) and Normalized Difference Moisture Index (NDMI)) were used as predictors of the models. Employing initial number of 82 presence and 250 absence data of D. littoralis, both models were able to predict the suitable areas for the species with fairly high success rate. The AUC and Kappa value for logistic regression were 0.77 ± 0.027 and 0.34 ± 0.058, respectively, while the respected values for Maxent were 0.91 ± 0.062 and 0.37 ± 0.025. Logistic regression analysis identified a total area of 26.13 km2 to be suitable for D. littoralis, while a smaller suitable area (7.85 km2) was predict- ed by Maxent model. Coastal areas in the west part of the island were predicted by both models as areas with high suitability for D. littoralis. Furthermore, distance from coastline and river, elevation, NDVI, NDWI and NDMI were suggested to be very important for the species ecology and distribution. The results of this study may serve as a basis for population reinforcement and reintroduction programs of D. littoralis and guide for ecosystem management of Nusakambangan Island as a whole. Key words: Cr itically endanger ed, Dipterocarpus littoralis, endemic species, logistic r egr ession, Maxent, Nusakambangan. ABSTRAK ROBIANSYAH, I. 2017. Prediksi distribusi habitat jenis endemik dan genting Dipterocarpus littoralis di Nusakambangan, Indonesia. Reinwardtia 16(1): 11 – 18. — Dipterocarpus littoralis (Blume) Kurz. adalah jenis endemik Nusakambangan dan termasuk ke dalam kategori genting. Pada penelitian ini kesesuaian habitat dari tumbuhan ini diprediksi menggunakan analisis regresi logistik dan model Maxent. Tiga variabel topografi (ketinggian, kelerengan dan arah lereng), jarak terhadap pantai dan sungai, serta satu index vegetasi Normalized Difference Vegetation Index (NDVI)) dan dua index kandungan air (Normalized Difference Water Index (NDWI) dan Normalized Difference Moisture Index (NDMI)) digunakan sebagai penduga untuk kedua model. Dengan menggunakan data awal berupa 82 titik keberadaan dan 250 titik ketidakhadiran jenis, kedua model ini dapat memprediksi kesesuaian habitat D. littoralis dengan tingkat kesuksesan yang cukup tinggi. Nilai AUC dan Kappa dari model regresi logistik secara berurutan ada- lah 0.77 ± 0.027 dan 0.34 ± 0.058, sedangkan bagi Maxent adalah 0.91 ± 0.062 dan 0.37 ± 0.025. Analisis regresi logistik memperkirakan total area yang sesuai bagi jenis ini adalah 26.13 km2 sedangkan model Maxent hanya seluas 7.85 km2. Kedua model ini memperlihatkan bahwa zona pesisir di daerah sebelah barat Nusakambangan adalah area dengan kesesuaian habitat yang sangat tinggi bagi D. littoralis. Selain itu, kedua model berhasil mengidentifikasi faktor lingkungan yang berpengaruh bagi ekologi dan persebaran spesies ini, yaitu jarak terhadap pantai dan sungai, ketinggian, NDVI, NDWI dan NDMI. Hasil dari penelitian ini dapat dijadikan dasar bagi program penguatan populasi alami dan reintroduksi D. littoralis dan sebagai panduan bagi pengelolaan ekosistem kawasan Nusakambangan secara keseluruhan. Kata kunci: Dipterocarpus littoralis, genting, jenis endemik, Maxent, Nusakambangan, regresi logistik. INTRODUCTION Small islands have very important role in the conservation of global plant biodiversity. Although they share only 5% of land surface of the earth, about one quarter of all known vascular plant species are endemic to islands (Kreft et al., 2008). As elsewhere, plant diversity of islands is under increasing pressure from habitat loss and conversion, population increase, introduction of invasive alien species, unsustainable use of native species, and climate change. Due to small geographic area, however, plant species in small islands are more sensitive to rapid environmental changes compared to other ecosystems. This eminent sensitivity of island plants is reflected in the high number of the island endemic extinction in the last 400 years (Sax & Gaines, 2008). The tree species Dipterocarpus littoralis (Blume) Kurz. is endemic to Nusakambangan, a small island located in Central Java, Indonesia. According to IUCN Red List 2015, the tree is categorized as REINWARDTIA 12 [VOL.16 Critically Endangered (Ashton, 1998). It is also included on a national list of priority species for conservation action in Indonesia 2008-2018. The major threats for the species are illegal cutting and the expansion in distribution of the Langkap tree (Arenga obtusifolia Mart.) (Robiansyah & Davy, 2015). To conserve and protect the species from extinction, several conservation measures have been implemented by government, local commu- nities, non government organization and/or univer- sity. These include information dissemination and public awareness program (USD, 2016), ecology and population status assessment (Robiansyah & Davy, 2015), population genetic research (Dwiyanti et al., 2014, Yulita & Partomihardjo, 2011), as well as propagation and reintroduction of D. littoralis (Holcim, 2013). To be successful, population reinforcement and reintroduction programs need to take into account habitat suitability of the target species. In case of D. littoralis, there is no detailed study assessing and describing suitable areas of the species in Nusakambangan. Previous modeling study assessing habitat suitability of D. litoralis by Primajati (2015) was incomplete and only covered the area of West Nusakambangan Nature Reserve. Thus the present study aims to assess habitat distribu- tion of D. littoralis in entire Nusakambangan Island. Using high resolution environmental variables and both presence-only and presence-absence modeling methods, the objectives of the present study are to: i) compare the ability of the two modeling methods in predicting habitat suitability of the species, ii) assess and describe suitable habitat areas of D. littoralis in Nusakambangan, and iii) identify environmental variables influencing the species distribution. The results of this study may serve as a basis for conservation programs of D. littoralis and the ecosystem management of Nusakambangan as a whole. MATERIALS AND METHODS Study site and occurrence data Nusakambangan is located in Indian Ocean, separated by a narrow strait off the southern coast of Central Java Province, Indonesia. It has an area of 210 km2 and the highest point at about 190 m above sea level. The climate is B type or Afa ac- cording to Koppen classification system. The eco- system is generally classified as lowland rainfor- est, and provides habitats for several rare and en- demic plant species, such as Rafflesia patma Blume, Lithocarpus platycarpus (Blume) Rehder, Gonystylus macrophyllus (Miq.) Airy Shaw, Anisoptera costata Korth., Shorea javanica Koord. & Valeton, Hopea sangal Korth., and A morphophal- lus decus-silvae Backer & Alderw. as well as Dip- terocarpus littoralis (Blume) Kurz (Partomihardjo et al., 2014). The endemic D. littoralis is a monoecious, emer- gent tree that grows up to 30 m tall. The wood is of good quality and is illegally harvested by local people for boat construction, timber, and firewood. In the present study, a total of 52 presence and 150 absence data of D. littoralis were obtained from the study conducted in West Nusakambangan Nature Reserve (WNNR) by Robiansyah and Davy (2015). The absence points were obtained by identifying all seemingly suitable habitats where the species was absent. In addition, 30 presence- only data were gathered from WNNR database and a total of 100 independent absence points were created randomly outside WNNR using ArcMap 10.1. Therefore, the present study initially used a total of 82 presence and 250 absence points for further modeling analysis. Environmental variables selection Elevation, slope and aspect were suggested to be very important for the ecology of D. littoralis (Robiansyah & Davy, 2015) and thus were select- ed as main predictors. Elevation data was obtained from Aster Global Digital Elevation Model V002 (http://earthexplorer.usgs.gov/) from which slope and aspect were derived using surface analysis extension in ArcMap10.1. Distance from river net- works and coastline were also used as predictors to see the effect of these two geographical features on the species distribution. In addition, one vegetation index (Normalized Difference Vegetation Index) and two water content indexes (Normalized Differ- ence Water Index and Normalized Difference Moisture Index) were used as a proxy of vegeta- tion greenness and soil moisture, respectively. All these indexes were derived from Landsat 8 OLI/ TIRS satellite image acquired in 6 May 2016 (http://earthexplorer.usgs.gov/) using the following formula (Sahu, 2014): Normalized Difference Vegetation Index (NDVI) = (Band 5 - Band 4) (Band 5 + Band 4) Normalized Difference Water Index (NDWI) = (Band 3 - Band 5) (Band 3 + Band 5) Normalized Difference Moisture Index (NDMI) = (Band 5 - Band 6) (Band 5 + Band 6) All the environmental layers used in the models had 30 m resolution. These layers were clipped to Nusakambangan boundary to be used in the modeling procedures. Modeling procedures Binary logistic regression and Maxent method were used to predict habitat suitability of D. litto- ralis using presence-absence and presence-only data, respectively. In binary logistic regression, all environmental variables intersected with 82 pres- ence and 250 absence points were extracted using ArcMap 10.1. The results were than analyzed using PASW Statistic 18 (SPSS Inc., 2009, ROBIANSYAH : Predicting habitat distribution of Dipterocarpus littoralis 2017] 13 www.spss.com) with aspect being converted into categorical variable to simplify the explanation of the results. Relative importance of each variable was estimated based on the absolute value of its standardized beta coefficient (Menard, 2011). For presence-only data, Maxent has been shown to perform better than other models (Elith et al., 2006) and has the best predictive power even with very low sample size (Wisz et al., 2008). It predicts the geographical distribution of a species using maximum entropy theory. In the present study, the Maxent software was set to the “auto feature”, logistic output format and ASCII output file type following the suggestion of Phillips and Dudík (2008). As Maxent automatically removes the duplicate records in the same cell, only 78 presence records were used in the modeling. Initially, a model was produced using Maxent that included all 8 environmental variables. Based on jackknife analysis provided by Maxent, variables contri- buting < 3% to the full model were excluded from the final model. To minimize the level of uncer- tainty and increase model accuracy, 10-fold cross- validation technique was used in which the final model was run 10 replicates and the results were then averaged. Furthermore, maximum training sensitivity plus specificity (MTSS) logistic thresh- old was used to convert the continuous suitability index into suitable and unsuitable areas of D. litto- ralis. This MTSS threshold is the best method for threshold selection when only presence data are available (Liu et al., 2013). The same threshold was also applied for the suitable areas distribution map resulted by logistic regression. The performance of both models was evaluated using the area under the curve (AUC) of the re- ceiver operating characteristic and Kappa value. AUC has been shown to be a highly effective threshold-independent measure of model perfor- mance (Thuiller et al., 2005) and its value varies from 0 (random prediction) to 1 (perfect predic- tion). For Kappa value, it is a robust statistic useful for model reliability testing, and its results can be range from -1 to +1 (McHugh, 2012). RESULTS Model performance The AUC and Kappa value for logistic regres- sion were 0.77 ± 0.027 and 0.34 ± 0.058, respec- tively, while the respected values for Maxent were 0.91 ± 0.062 and 0.37 ± 0.025. These values indi- cated the fairly high success rate of both models in predicting potential distribution of D. littoralis. Araújo et al. (2005) suggested the following standard for judging model performance based on AUC value: excellent (AUC>0.9), good (0.9>AUC>0.8), fair (0.8>AUC>0.7), poor (0.7>AUC>0.6), and failed (0.6>AUC>0.5). For Kappa statistic, its values can be interpreted as follows: values ≤ 0 as indicating no agreement and 0.01–0.20 as none to slight, 0.21–0.40 as fair, 0.41– 0.60 as moderate, 0.61–0.80 as substantial, and 0.81–1.00 as almost perfect agreement (McHugh, 2012). Contribution of environmental variables Logistic regression analysis identified four en- vironmental variables being significantly influen- tial (χ2=46.45, df= 4, P<0.000) for D. littoralis distribution (Table 1). These factors include NDWI, NDMI, distance from coastline and eleva- tion. The first factor had the highest influence to the model as indicated by its high absolute standardized beta value. In addition to these four factors, Maxent predicted that distance from river networks and NDVI were also important for the species (Table 2). The model also revealed that distance from coastline was the main determinant factor of D. littoralis distribution in Nusa- Variable Absolute standardized beta P Normalized Difference Water Index 0.31 0.000 Normalized Difference Moisture Index 0.17 0.027 Distance from coastline 0.10 0.017 Elevation 0.09 0.000 Variable Percent contribution Distance from coastline 36.4 Elevation 21.8 Normalized Difference Vegetation Index 17 Distance from river networks 16.4 Normalized Difference Moisture Index 8.4 Table 1. Environmental variables significantly contribute to the presence of Dipterocarpus littoralis in Nusakambangan based on logistic regression analysis. Table 2. Relative contribution (%) of the environmental variables to the Maxent model of Dipterocarpus littoralis REINWARDTIA 14 [VOL.16 kambangan with more than 36% contribution to the model. Response curves of environmental variables produced by Maxent model (Fig. 1) showed the specific environmental preferences of D. litto- ralis. The species had high probability of being presence at areas with distance of 200-750 m from coastline and 300-900 m from rivers, elevation of 30-75 m above sea level, NDVI of 0.75-0.79 and NDMI above 0.45. Habitat distribution of D. littoralis A total area of 26.13 km 2 was predicted by logistic regression analysis to be suitable for D. littoralis . These areas were distributed mainly in the west and north part of the island (Fig. 2). A smaller suitable area (7.85 km ) was predicted by 2 Maxent model and found mainly in West Nusa- kambangan coastal areas. Furthermore, as it was agreed by both logistic and Maxent, the intersec- tion areas of both models were predicted to have very high suitability for the species. These areas had an area of 3.82 km2 and were located mainly in the west part of the island. DISCUSSION Using logistic regression and Maxent model, the present study was able to predict the distribution of suitable areas for D. littorralis with high success rate. The Maxent model, however, outperformed the logistic regression method as showed by its Fig. 1. Response curves showing the relationship between the probability of presence of Dipterocarpus littoralis and the significant environmental variables. Values shown are average over 10 replicate runs; blue margins show ±1 SD calculated over 10 replicates. ROBIANSYAH : Predicting habitat distribution of Dipterocarpus littoralis 2017] 15 higher AUC and Kappa value. This superiority of Maxent model compared to the logistic regression was also shown by previous studies (Gastón & García-Viñas, 2011; Marini et al., 2010; Tognelli et al., 2009; Roura-Pascual et al., 2009), and main- ly due to its ability to ovoid overfitting by using regularization techniques (Gastón & García-Viñas, 2011). In addition, the AUC value of Maxent model from the present study was much higher compared to the AUC of 0.87 gained by Primajati (2015). In general the results of the present study were in concordance with the study of Robiansyah and Davy (2015) and Primajati (2015), which high- lighted the importance of elevation, distance from coastline and river networks, soil moisture and canopy cover for the ecology and distribution of D. littoralis. Interesting results were shown by logistic regression and Maxent model which identified NDWI and distance from coastline as the most important variable for D. littoralis distribution, respectively. NDWI has a strong positive correla- tion with fractional water index (Gu et al., 2008), thus is a very good proxy for soil moisture condi- tion. In the present study, predicted suitable habitat for D. littoralis was associated with NDWI value above 0.35 (result was not shown), which corre- sponds to high soil moisture content. As previous study had shown that more than 60% of D. litto- ralis in WNNR were in 0-5 cm diameter class (Robiansyah & Davy, 2015), this high soil mois- ture content might be very important for seedling growth and establishment of D. littoralis. A study by Islam et al. (2016) found a positive association between sapling density of D. turbinatus and soil moisture content in Lawachara National Park, Bangladesh. The authors suggested that soil mois- ture content is a critical factor for the regeneration and a viable population of the species. As indicated by its name, D. littoralis is an en- demic tree found in littoral forest of Nusa- kambangan. A littoral forest is recognized by its close proximity to the ocean (generally < 2 km) (DECC, 2008). This might be the reason Maxent identified distance to coastline as the most im- portant factor for the distribution of D. littoralis. Furthermore, inclusion of this variable as a signifi- cant predictor in the Maxent model restricted the distribution of these habitats mainly along the coastal areas of the island, and hence led to the prediction of smaller suitable habitats of the spe- cies compared to those predicted by logistic re- gression. Most of the intersection areas between predict- ed suitable habitats of logistic regression and Maxent model were located inside WNNR. Since these intersection areas were predicted to have very high suitability for D. littoralis, this situation is good for the species in term of conservation strategies as these areas are already protected by the government. Furthermore, the undergoing and future population reinforcement and reintroduc- tion of the species should be focused in these in- tersection areas in order to increase the successful- ness of the programs. Hai et al. (2014) argued that site selection, habitat features and geographical location are among the most important factors affecting the success of plant reintroduction program. Since there were some limitations in the present modeling study, care should be taken when imple- menting the results in the field-based conservation programs. Logistic regression and Maxent model did not consider other important factors for the distribution of a plant species such as dispersal process, anthropogenic influences, biotic interac- tions, or geographic barriers (Pearson, 2010; Soberón, 2007). In addition, microclimate variables were also not included in both models since there was no weather station in the island from which Fig. 2. Predicted distribution of Dipterocarpus littoralis in Nusakambangan Island based on logistic regression analysis (yellow), Maxent model (green) and intersection areas of the two models (red). REINWARDTIA 16 [VOL.16 temperature and precipitation related variables can be collected. Furthermore, field surveys are re- quired to validate the model predictions, especial- ly outside WNNR where the presence and absence data of D. littoralis were absent. CONCLUSION The present study showed that Maxent model was better than logistic regression analysis in pre- dicting the suitable habitats for D. littoralis. The AUC of the first model was much higher com- pared to the second one. Distance from coastline and river networks NDWI, NDVI and elevation were suggested to be very important for the spe- cies ecology and distribution. Coastal areas in the west part of the island were predicted by both models as areas with high suitability for D. litto- ralis. Although there were some limitations, the results of the present study can be used as a basis for conservation programs of D. littoralis. Popula- tion reinforcement and reintroduction programs can be focused at the areas with high suitability for the species, especially at the intersection areas predicted by both logistic regression and Maxent model. Furthermore, these predicted suitable areas should be protected as they may serve as in situ refugia for D. littoralis from ongoing climate change. REFERENCES ARAÚJO, M. B., PEARSON, R. G., THUILLER, W. & ERHARD, M. 2005. Validation of species –climate impact models under climate change. Glob. Change Biol.11: 1504–1513. ASHTON, P. 1998. Dipterocarpus littoralis. The IUCN Red List of Threatened Species 1998: e.T33376A9773710.http://doi.org/10.2305/IUCN. UK.1998.RLTS.T33376A9773710.en.Accessed 29 August 2016 DECC (Department of Environment & Climate Change). 2008. Littoral Rainforest in the South East Corner, Sydney Basin and NSW North Coast bioregions. http://www.environment. nsw.gov.au/resources/threatenedspecies / EEClittoralrainforestlowres.pdf Accessed 29 August 2016. DWIYANTI, F. G., HARADA, K., SIREGAR, I. Z. & KAMIYA, K. 2014. Population genetics of the critically endangered species Dipterocarpus littoralis Blume (Dipterocarpaceae) endemic on Nusakambangan Island, Indonesia. Biotropia 21: 1-12. ELITH, J., GRAHAM, C. H., ANDERSON, R. P., DUDIK, M., FERRIER, S., GUISAN, A., HIJMANS, R. J., HUETTMANN, F., LEATHWICK, J. R., LEHMANN, A., LI, J., LOHMANN, L. G., LOISELLE, B. A., MAN- ION, G., MORITZ, C., NAKAMURA, M., NAKAZAWA, Y., OVERTON, J. M. M., PE- TERSON, A. T., PHILLIPS, S. J., RICHARD- SON, K., SCACHETTI-PEREIRA, R., SCHAPIRE, R. E., SOBERO´N, J., WIL- LIAMS, WISZ, S. M. & ZIMMERMANN, N. E. 2006. Novel methods improve prediction of species’ distributions from occurrence data. Ecography 29: 129–151. GASTÓN, A. & GARCÍA-VIÑAS, J. I. 2011. Modelling species distributions with penalised logistic regressions: A comparison with maxi- mum entropy models. Ecol. Model. 222: 2037– 2041. GU, Y., HUNT, E., WARDLOW, B., BASARA, J. B., BROWN, J. F. & VERDIN, J. P. 2008. Eval- uation of MODIS NDVI and NDWI for vegeta- tion drought monitoring using Oklahoma Mesonet soil moisture data. Geophys. Res. Lett. 35, L22401. HAI, R., SHUGUANG, J., HONGXIAO, L., QI- ANMEI, Z., & HONGFANG, L. 2014. Advanc- es in the reintroduction of rare and endangered wild plant species. Sci. China Life Sci. 57(6): 603–609. HOLCIM. 2013. Kinerja Pengelolaan Lingkungan Juli 2012-Juni 2013. http://proper.menlh.go.id/ portal/filebox/DRKPL%202013%20HOLCIM% 20CILACAP.pdf Accessed 27 June 2016. ISLAM, M., SALIM, S. H., KAWSAR, M. H. & RAHMAN, M. 2016. The effect of soil moisture content and forest canopy openness on the re- generation of Dipterocarpus turbinatus CF Gaertn.(Dipterocarpaceae) in a protected forest area of Bangladesh. Trop. Ecol. 57: 455–464. KREFT, H., JETZ, W., MUTKE, J., KIER, G. & BARTHLOTT, W. 2008. Global diversity of island floras from a macroecological perspec- tive. Ecol. Lett. 11: 116–127. LIU, C., WHITE, M. & NEWELL, G. 2013. Selecting thresholds for the prediction of species occurrence with presence-only data. J. Biogeogr. 40: 778–789. MARINI, M. Â., BARBET-MASSIN, M., LOPES, L. E. & JIGUET, F. 2010. Predicting the occur- rence of rare Brazilian birds with species distri- bution models. J. Ornithol. 151: 857–866. MCHUGH, M.L. 2012. Interrater reliability: the kappa statistic. Biochem. Med. 22(3): 276 -282. MENARD, S. 2011. Standards for standardized logistic regression coefficients. Soc. Forces 89 (4): 1409–1428. PARTOMIHARDJO, T., ARIFIANI, D., PRATA- MA, B.A. & MAHYUNI, R. 2014. Jenis Jenis Pohon Penting di Hutan Nusakambangan. LIPI Press, Jakarta. PEARSON, R. G. 2010. Species distribution modeling for conservation educators and practitioners. Lessons Conserv. 3: 54–89. ROBIANSYAH : Predicting habitat distribution of Dipterocarpus littoralis 2017] 17 PHILLIPS, S. J. & DUDÍK, M. 2008. Modeling of species distributions with Maxent: new exten- sions and a comprehensive evaluation. Ecogra- phy 31: 161–175. PRIMAJATI, M. 2015. Prediksi Kesesuaian Habi- tat Plahlar (Dipterocarpus littoralis) di Kawa- san Cagar Alam Nusa Kambangan Barat. Flora Fauna Indonesia, Jakarta. ROBIANSYAH, I. & DAVY, A. J. 2015. Popula- tion status and habitat preferences of critically endangered Dipterocarpus littoralis in West Nu- sakambangan, Indonesia. Makara J. Sci. 19: 150 –160. ROURA-PASCUAL, N., BROTONS, L., PETER- SON, A. T. & THUILLER, W. 2009. Consensu- al predictions of potential distributional areas for invasive species: a case study of Argentine ants in the Iberian Peninsula. Biol. Invas. 11: 1017– 1031. SAHU, A. S. 2014. Identification and mapping of the water-logged areas in Purba Medinipur part of Keleghai river basin, India: RS and GIS methods. Int. J. A dv. Geosci. 2(2): 59–65. SAX, D. F. & GAINES, S. D. 2008. Species inva- sions and extinction: The future of native biodi- versity on islands. P. Natl. A cad. Sci. 105: 11490–11497. SOBERÓN, J. 2007. Grinnellian and Eltonian niches and geographic distributions of species. Ecol. Lett. 10: 1115–1123. THUILLER, W., LAVOREL, S. & ARAU´JO, M. B. 2005. Niche properties and geographical ex- tent as predictors of species sensitivity to cli- mate change. Glob. Ecol. Biogeogr. 14: 347– 357. TOGNELLI, M. F., ROIG-JUNENT, S. A., MARVALDI, A. E., FLORES, G. E. & LOBO, J. M. 2009. An evaluation of methods for modelling distribution of Patagonian insects. Rev. Chil. Hist. Nat. 82: 347-360. USD. 2016. Bincang Plahlar di PSL USD. Humas Universitas Sanata Dharma. https://www.usd.ac. id/deskripsi.php?idt=usd_berita&noid=2672 Accessed 27 June 2016. WISZ, M. S., HIJMANS, R. J., LI, J., PETER- SON, A. T., GRAHAM, C. H., GUISAN, A. & GROUP, N. P. S. D. W. 2008. Effects of sam- ple size on the performance of species distribu- tion models. Divers. Distrib. 14: 763–773. YULITA, K. S. & PARTOMIHARDJO, T. 2011. Keragaman genetika populasi Pelahlar (Dipterocarpus littoralis (Bl.) Kurz) di Pulau Nusakambangan berdasarkan Profil Enhanced Random Amplified Polymorphic DNA. Berita Biologi 10: 541–548. REINWARDTIA 18 [VOL.16 INSTRUCTION TO AUTHORS Scope. R einwardtia is a scientific ir r egular jour nal on plant taxonomy, plant ecology and ethnobotany published in June and December. Manuscript intended for a publication should be written in English. Titles. Titles should be br ief, infor mative and followed by author ’s name and mailing address in one- paragraphed. Abstract. English abstr act followed by Indonesian abstr act of not mor e than 250 wor ds. Keywor ds should be given below each abstract. Manuscript. Manuscr ipt is or iginal paper and r epr esent an ar ticle which has not been published in any other journal or proceedings. The manuscript of no more than 36 pages by using Times New Roman 11, MS Word for Windows of A4 with double spacing, submitted to the editor through . New paragraph should be indented in by 5 characters. For the style of presentation, authors should follow the latest issue of Reinwardtia very closely. Author(s) should send the preferred running title of the article submitted. Every manuscript will be sent to two blind reviewers. Identification key. Taxonomic identification key should be pr epar ed using the aligned couplet type. Nomenclature. Str ict adher ence to the Inter national Code of Nomenclatur e is obser ved, so that taxonomic and nomenclatural novelties should be clearly shown. English description for new taxon proposed should be provided and the herbaria where the type specimens area deposited should be presented. Name of taxon in taxonomic treatment should be presented in the long form that is name of taxon, author’s name, year of publication, abbreviated journal or book title, volume, number and page. Map/line drawing illustration/photograph. Map, line dr awing illustr ation, or photogr aph pr efer ably should be prepared in landscape presentation to occupy two columns. Illustration must be submitted as original art accompanying, but separated from the manuscript. The illustration should be saved in JPG or GIF format at least 350 pixels. Legends or illustration must be submitted separately at the end of the manuscript. References. Bibliogr aphy, list of liter atur e cited or r efer ences follow the Har var d system as the following examples. Journal : KRAENZLIN, F. 1913. Cyrtandraceae novae Philippinenses I. Philipp. J. Sci. 8: 163–179. MAYER, V., MOLLER, M., PERRET, M. & WEBER, A. 2003. Phylogenetic position and generic differentiation of Epithemateae (Gesneriaceae) inferred from plastid DNA sequence data. American J. Bot. 90: 321–329. Proceedings :TEMU, S. T. 1995. Peranan tumbuhan dan ternak dalam upacara adat “Djoka Dju” pada suku Lio, Ende, Flores, Nusa Tenggara Timur. In: NASUTION, E. (Ed.). Prosiding Seminar dan Lokakarya Nasional Etnobotani II. LIPI & Perpustakaan Nasional: 263–268. (In Indonesian). SIMBOLON, H. & MIRMANTO, E. 2000. Checklist of plant species in the peat swamp forests of Central Kalimantan, Indonesia. In: IWAKUMA, T. et al. (Eds.) Proceedings of the International Symposium on: Tropical Peatlands. Pp.179-190. Book : RIDLEY, H. N. 1923. Flora of the Malay Peninsula 2. L. Reeve & Co. Ltd, London. Part of Book : BENTHAM, G. 1876. Gesneriaceae. In: BENTHAM, G. & HOOKER, J. D. Genera plantarum 2. Lovell Reeve & Co., London. Pp. 990–1025. Thesis : BAIRD, L. 2002. A Grammar of Kéo: An Austronesian language of East Nusantara. Australian National University, Canberra. [PhD. Thesis]. Website : http://www.nationaalherbarium.nl/fmcollectors/k/KostermansAJGH.htm). Accessed 15 February 2012. http://www.nationaalherbarium.nl/fmcollectors/k/KostermansAJGH.htm REINWARDTIA Author Agreement Form Title of article : Name of Author(s) : I/We hereby declare that:  My/Our manuscript was based on my/our original work.  It was not published or submitted to other journal for publication.  I/we agree to publish my/our manuscript and the copyright of this article is owned by Reinwardtia.  We have obtained written permission from copyright owners for any excerpts from copyrighted works that are included and have credited the sources in our article. Author signature (s) Date ___________________________________________________________________________________________ Name Reinwardtia Published by Herbarium Bogoriense, Botany Division, Research Center for Biology, Indonesian Institute of Sciences Address: Jln. Raya Jakarta-Bogor Km. 46 Cibinong 16911, P.O. Box 25 Cibinong Telp. (+ 62) 21 8765066; Fax (+62) 21 8765062 E-mail: reinwardtia@mail.lipi.go.id REINWARDTIA Vol. 16. No. 1. 2017 CONTENTS Page DINI PUSPITANINGRUM, WENDY A. MUSTAQIM & MARLINA ARDIYANI. A new record of Etlingera pauciflora (Zingiberaceae) in Java, Indonesia ...…………………………………..…………………..…….…...………….…………. 1 SRI RAHAYU & MICHELE RODDA. Hoya narcissiflora (Apocynaceae, Asclepiadoideae), a new species from Borneo ……………………………………………………………………………………….…………………………. 5 IYAN ROBIANSYAH. Predicting habitat distribution of endemic and critically endangered Dipterocarpus littoralis in Nusakambangan, Indonesia ……………………………………………….…..……………………………….………. 11 LULUT DWI SULISTYANINGSIH. A newly described and recorded infraspecific taxa of Musa borneensis Becc. (Musaceae) from Sulawesi, Indonesia ……………………………………………...…....…………….………………..... 19 JAN-FIRTS VELDKAMP & ABDULROKHMAN KARTONEGORO. New species of Catanthera and Medinilla (Melastomataceae) from Halmahera, Indonesia and a new name for a Medinilla from Madagascar…………………………………...…...……... 25 PURWANINGSIH, RUDDY POLOSAKAN, RAZALI YUSUF & KUSWATA KARTAWINATA. Phytosociological study of the montane forest on the south slope of Mt. Wilis, East Java, Indonesia………………………..…………………………….…. 31 IAN M. TURNER. A new combination for subspecies of Radermachera quadripinnata (Bignoniaceae) ........................ 47 Reinwardtia is a LIPI accredited Journal (792/AU3/P2MI-LIPI/04/2016) http://e-journal.biologi.lipi.go.id/index.php/reinwardtia Herbarium Bogoriense Botany Division Research Center for Biology – Indonesian Institute of Sciences Cibinong Science Center Jln. Raya Jakarta − Bogor, Km 46 Cibinong 16911, P.O. Box 25 Cibinong Indonesia front cover, cover dalam, reviewer Reinwardtia 16(1)_1-1 REINWARDTIA 2017 Jun Full_2-2 REINWARDTIA 2017 Jun Full_3-3 REINWARDTIA 2017 Jun Full_4-4 REINWARDTIA 2017 Jun Full_15-15 REINWARDTIA 2017 Jun Full_16-16 REINWARDTIA 2017 Jun Full_17-17 REINWARDTIA 2017 Jun Full_18-18 REINWARDTIA 2017 Jun Full_19-19 REINWARDTIA 2017 Jun Full_20-20 REINWARDTIA 2017 Jun Full_21-21 REINWARDTIA 2017 Jun Full_22-22 8. Daftar Isi Reinwardtia 16(1)