Mapping of biodiversity hubs and key ecosystem services as a tool for shaping optimal areas for conservation

often-heard challenge is that not all ES can be mapped, the most frequently mapped are provi- sioning ES (food, timber, water, etc.) and regulating ES (climate regulation) but cultural ES are less so [60]. As for supporting services, they are necessary ES for the production of all other ES. Some examples include biomass production, production of atmospheric oxygen, soil forma- tion and retention, nutrient cycling, water cycling, and provisioning of habitat [61]. In developing world countries where data gaps are a major challenge the set of ES possible to analyze is generally constrained by the availability and consistence of spatial data at national scale, the cost, processing time, and level of expertise. Tools used in our investigation include Co$ting Nature (CN), Water World (WW) and ESRI’s ArcGIS 10.2.2. For a complete review of ES assessment tool, please see [62–65], forCN and WW, and [66] for ArcGIS 10.2.2. WW/ CN are particularly attractive with the simplicity of the tools and wealth of ES related data available with them. Among their strengths over other site-scale ES tools is that they are web- based, freely available to users, does not require special GIS expertise, complete analysis in less time, have 1ha to 1 km2 spatial resolution, have monthly/diurnal temporal resolution, supply required biophysical data, provide data for climate change impacts, land use change impacts, land management impacts, water management impacts scenarios, and provide results about baseline services such as water quality, water quantity, water regulation, carbon, etc.

In this study, we assessed four key regulating ES: Carbon storage, Water balance, Human footprint on water quality/diarrheal disease and Human footprint on water quality/annual. These services were selected based on their significance to Rwandan context and the availabil- ity of data. All spatial analysis was conducted and maps generated respectively using ESRI’s ArcGIS 10.2.2 [66], CN and WW. 2.5.1 Carbon storage. Carbon storage refers to the amount of carbon seized by terrestrial ecosystems (both aboveground and belowground) [9], the major carbon reservoirs are esti- mated to be oceans (38000 Pg C), terrestrial vegetation (500 Pg C), soils (1500–2000 Pg C), fos- sil fuels (5000–10000 Pg C) and atmosphere (780 Pg C); the exchanges of carbon between the atmosphere and terrestrial ecosystems is critical to the patterns of CO2 concentration in the atmosphere [67]. The natural stock of carbon benefits humans by reducing or stabilizing the atmospheric greenhouse gas concentration and mitigate climate change [58]. By considering different land use/land cover the carbon storage in Rwanda’s natural forest, non-forested, wet- land and water bodies the carbon storage was estimated and its map was drawn. We downloaded 4 satellite images taken by Landsat 8 OLI/TIRS from the U.S. Geological Survey Earth Explorer, tool that gives users the ability to query, search, and order aerial photo- graphs, satellite images and cartographic products. The images were subtracted from the fol- lowing address: Path 172 Row 61; Path 172 Row 62; Path 173 Row 61; Path 173 Row 62 and

the date range was from 01/01/2019 up to 31/12/2019. Images with less than 10% of cloud all captured during the same dry season were preferred to minimize natural landscape elements fluctuations related to seasonal change. The most recent Landsat 8 mission carries the Opera- tional Land Image (OLI) and the Thermal Infrared Sensor (TIRS). OLI/TIRS data are delivered at 30 m resolution with 11 spectral bands. The images are georeferenced and hold many layers containing data collected along both visible and invisible light spectrum that can be manipu- lated to reveal hidden indications about pictured ground features as highlighted in Table 1. The richness of Landsat archive combined with a no cost data policy, gave users an unequalled opportunity to monitor the natural and human-induced changes on the global landscape [68]. We have performed radiometric and geometric calibration on downloaded images to cor- rect and eliminate distortions caused by the Earth’s rotation and camera angles. We have com- bined the multiple raster datasets into a single Mosaic image on which we have performed the supervised land use/cover classification, the common technic of remote sensing in which the image processing software (ArcGIS) is given primary guidance by the operator to specify the PLOS ONE How do old national parks cope with regenerating conservation mission? PLOS ONE | https://doi.org/10.1371/journal.pone.0253151 August 17, 2021 7 / 20