Practice makes the expert: The importance of training volunteers in the generation of phenological data from photographs of biodiversity observation platforms

Mexico have phenological annotations (data obtained from naturalista.mx), perhaps due to a lack of obligation to record this information, or lack of interest or knowledge about phenologi- cal recording. Although these platforms are not explicitly aimed at generating data on phenol- ogy, by including photographs they provide a very useful visual verification resource for researchers that allows for since it allows corroboration of the species identity, the pheno- phases that occur, and the state of cultivation, which is often a problem with citizen science observations without photographic records [24,41]. The aim of this study was to use Naturalista, a biodiversity observation platform with photo- graphic resources, as an input of phenological information at a national scale in Mexico to evaluate the benefits and limitations of this information source in the generation of phenological informa- tion in a wide geographical area. Since some of these platforms are not specifically designed to cap- ture phenological data, one of the specific goals was to propose a methodology to reliably obtain and classify phenological information from photographic records of species presence, with the participation of citizen scientists. We assessed whether the quality of phenological classification by volunteers or citizen scientists changed after a brief but concise training on basic biology and phe- nology, and compared the level of agreement in phenological classification among groups of expert scientists, non-expert volunteers with training, and non-expert volunteers without training.

Methods This study was conducted using three components: (1) a photographic database of the Natura- lista digital platform (https://www.naturalista.mx/), where photographs were classified accord- ing to the phenological phases they represented; (2) a group of five ecologists who have field experience with the studied species (hereafter referred to as a group of experts), who defined the classification criteria for the photographic records—the phenophases—characteristics of eachphase, and possible ratings; and (3) a group of 49 citizen scientists (hereafter called volun- teers) who were divided into two groups: with and without training, for the subsequent classifi- cation of the photographs. Two species of invasive plants in Mexico were used to develop the proposed methodology and identify the potential difficulties with it (Fig 1). Since the methodology focused on generat- ing phenological data at a regional or larger scale, we select species that have a wide distribu- tion, and a large number of records to evaluate phenological information over one year. Invasive species, although they are not the only species that have these characteristics, nor are they necessarily the species with the highest number of observations in Naturalista (www. naturalista.mx), provide good case studies because, they offer large amounts of data over a large geographical area. The presence of clear phenological phases was an important criterion, so the selected invasive species had conspicuous reproductive and vegetative structures (i.e.,

with considerable size as well as a vibrant color and shape easily distinguished from the rest of the plant). Two species that fulfilled the above criteria were selected for the study: Leonotis nepetifolia (L.) R. Brown (1,234 records) and Nicotiana glauca Graham (1,583 records). For both species, four easily identifiable phenological stages were defined: (1) initial growth, which corresponds to the presence of leaf buds and young leaves identified by their shape, size, and color; (2) immature flower, which corresponds to flower buds; (3) mature flower, which includes flowers with and unfolded corolla; and (4) dry fruits. Once the phenophases were selected, the group of experts designed an informative guide with the main characteristics of the plant structures that define each phenophase for each spe- cies, detailing the specific structures to be identified in each phenophase and the scoring proto- col. Not all phenophases and structures were easily identifiable through photographs. For example, for L. nepetifolia, the immature fruit remains within one of the bracteoles that are PLOS ONE The training of volunteers as a promoter of phenological studies PLOS ONE | https://doi.org/10.1371/journal.pone.0282750 March 7, 2023 4 / 17