LETTER • OPEN ACCESS You may also like - Recent changes in phenology over the Aligning citizen science and remote sensing northern high latitudes detected from multi- satellite data phenology observations to characterize climate Heqing Zeng, Gensuo Jia and Howard Epstein change impact on vegetation - Observed earlier start of the growing season from middle to high latitudes across the Northern Hemisphere snow- To cite this article: Cristina Domingo-Marimon et al 2022 Environ. Res. Lett. 17 085007 covered landmass for the period 2001–2014 Xiaona Chen and Yaping Yang - Modeling phenological responses of Inner Mongolia grassland species to regional View the article online for updates and enhancements. climate change Qiuyue Li, Lin Xu, Xuebiao Pan et al. This content was downloaded from IP address 158.109.186.12 on 01/02/2024 at 10:59 Environ. Res. Lett. 17 (2022) 085007 https://doi.org/10.1088/1748-9326/ac8499 LETTER Aligning citizen science and remote sensing phenology OPEN ACCESS observations to characterize climate change impact on vegetation RECEIVED 9 January 2022 Cristina Domingo-Marimon1,∗, Joan Masó1, Ester Prat1, Alaitz Zabala2, Ivette Serral1, REVISED Meritxell Batalla1, Miquel Ninyerola3 and Jordi Cristóbal4,5 22 July 2022 1 ACCEPTED FOR PUBLICATION Grumets Research Group, CREAF, Edifici C, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès, Catalunya, Spain 2 27 July 2022 Grumets Research Group, Geography Department, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Catalunya, 08193, Spain 3 Grumets Research Group BABVE, Universitat Autònoma de Barcelona, Cerdanyola del Vallès 08193, Spain PUBLISHED 4 8 August 2022 Efficient Use of Water in Agriculture Program, Institute of Agrifood Research and Technology, Fruitcentre, Parc Científic i Tecnològic Agroalimentari de Lleida 23, 25003 Lleida, Spain 5 Geophysical Institute, University of Alaska Fairbanks, 903 Koyukuk Dr., Fairbanks, AK 99775-7320, United States of America Original content from ∗ Author to whom any correspondence should be addressed. this work may be used under the terms of the E-mail: cristina.domingo@uab.cat Creative Commons Attribution 4.0 licence. Keywords: citizen science, phenology, phenometrics, Sentinel-2, time-series, remote sensing Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal Abstract citation and DOI. Phenology observations are essential indicators to characterize the local effects of climate change. Citizen participation in the collection of phenological observations is a potential approach to provide data at both high temporal scale and fine grain resolution. Traditional observation practices of citizen science (CS), although precise at the species scale, are limited to few observations often closely located to an observer’s residence. These limitations hinder coverage of the great variability of vegetation phenology across biomes and improvement of the knowledge of vegetation changes due to climate change impacts. This study presents a new approach to overcome these limitations by improving CS guidance and feedback as well as expanding phenology report sites and observations across different habitats and periods to contribute to monitoring climate change. This approach includes: (a) a new methodology focused on harmonizing remote sensing phenology products with traditional CS phenology observations to direct volunteers to active phenology regions and, (b) a new protocol for citizen scientists providing tools to guide them to specific regions to identify, collect and share species phenological observations and their phenophases. This approach was successfully tested, implemented and evaluated in Catalonia with more than 5000 new phenologically interesting regions identified and more than 200 observations collected and Sentinel-2 derived phenometrics were demonstrated as of good quality. 1. Introduction increase efficiency of traditional research (McKinley et al 2017). Indeed, observational data collected by In the last decade, public participation in environ- citizen science (CS) is considered a valuable data mental research, understood as any voluntary contri- source in many environmental analyses related to bution of citizens engaged in the process of scientific monitoring ecosystem activities (Kirchhoff et al 2021, research to generate new science-based knowledge Pucino et al 2021), analyses of trends and drivers or (Shirk et al 2012, Fraisl et al 2020), has grown rapidly nature conservation (García et al 2021). Moreover, in (MacPhail and Colla 2020). Contributions to science the last decades CS projects have grown in number and conservation are identified as the most import- and scope (Burgess et al 2017). ant motivations for participation and often become Further, scientifically managed CS projects more significant as project participation progresses provide the valuable asset of bringing research closer (Larson et al 2020). to the general public. This improves society’s per- Public participation has also been properly recog- ception of the value of science by amplifying under- nized as a potentially useful tool to reduce costs and standing of the subject matter (Forrester et al 2017) © 2022 The Author(s). Published by IOP Publishing Ltd Environ. Res. Lett. 17 (2022) 085007 C Domingo-Marimon et al and increasing trust between scientists and the gen- Crimmins et al 2022), although there are some quite eral public (Bonney et al 2014), a challenge that the well established programs working in this direction, scientific community has been historically pursuing. such as the USA National Phenology Network or the Ground observations made by the general public Pan European Phenology Project PEP725. On the are not something new. Traditionally, farmers or nat- other hand, the variability of vegetation phenology uralists and some scientists have recorded environ- across biomes is significantly large (Richardson et al mental observations for decades in their daily work 2013), so having more data to improve the knowledge or as a hobby, ranging from meteorological data to about these changes is a must and a challenge that can biodiversity sightings, and some of them also record- hardly be fulfilled by volunteers (Bonney et al 2014, ing phenology. Phenology, understood as the phases Morrison 2016, Chandler et al 2017). Unfortunately, of the life cycle of living organisms (plants or anim- currently, there are not many other projects involving als) and how they are affected by seasonal and interan- CS and phenology monitoring that could guide the nual climate variations, has been a common societal volunteers (MacPhail and Colla 2020) and increase topic (White et al 1997, de Beurs and Henebry 2004), the number of participants. and is booming in the context of climate change, This context raises a new challenge for scient- as such observations have become an indisputable ists wishing to improve participation in environ- indicator and regulator of the impacts of climate mental research, who have to build guidance to volun- change (Peñuelas et al 2009, Richardson et al 2013, teers, assess the quality of collected data, and expand Scranton and Amarasekare 2017). In fact, changes the observed sites across different biomes or hab- in phenology corresponding to specific vegetation itats (McDonough MacKenzie et al 2017). Indeed, physiological processes (leaf unfolding and flower- meeting this challenge will have to involve differ- ing of plants in spring, fruit ripening, colour chan- ent methodologies combining citizen-based observa- ging and leaf fall in autumn, among others) led to tions and technological methods to manage limiting adaptation to new environmental resources (Feldman factors related to volunteers’ participation and large- et al 2018, Zeng et al 2020). Also, changes in his- scale and long-term temporal and spatial coverage. torical seasonality of vegetation phenology are recog- With regard to the latter constraints, Earth obser- nized as a sensitive indicator of climate change and its vation from space has become a highly used techno- local effects in terms of ecosystem response (Ganguly logy to observe global change phenomena at a large et al 2010). However, frequent and repeated observa- scale, thanks to the amount of satellites in orbit con- tions over time are required to perceive these changes stantly revisiting the Earth (Duchemin et al 1999, (Feldman et al 2018). White et al 2009, Nagai et al 2010, Isaacson et al 2012, Although CS is considered a potential source of Evans et al 2014). Thus, remote sensing (RS) provides data at a high temporal scale and fine grain resolution effective means to observe global change with tem- (Burgess et al 2017) so far, ground-observed phen- poral and spatial precision (Guo 2014). Undoubtedly, ology, although precisely at species scale, is reduced long time spatially continuous resolutions of RS data to small traditional elements closely located to an are essential characteristics for phenology monitor- observer’s home (White et al 2005). Generally speak- ing because they capture in detail the monitored ing, volunteers’ efforts may be biased by represent- seasonal variation in vegetation dynamics. Current, ing a small demarcated area but in a very precise RS products present too coarse a view of pheno- and systematic way, as they are usually more likely logy dynamics, insensitive to small changes in spe- to collect data from accessible locations (close to cies and habitat distribution. The lack of focus on the roads or urban areas) (Dickinson et al 2010, individual species or habitats creates two problems. Phillimore et al 2013, Crimmins et al 2022). Never- Firstly, since each species has different sensitivity to theless, through worldwide observers, this is a mod- climate change (Thuiller 2004, Schuetz et al 2019), est but not a negligible approximation to cover the it is difficult to assess the effects of climate change great variability of vegetation phenology across bio- on them with available RS products. Secondly, since mes or habitats aiming at improving the knowledge the phenology observations by CS look to individual of vegetation changes as a consequence of climate plants in detail, it is not easy to evaluate the medium change impacts. If the strategy is to be comprehens- resolution RS products against in-situ measurements. ive and systematic, this ground-based approach faces Current medium spatial resolution satellite sensors, major gaps. On the one hand, volunteers particip- such as the Moderate Resolution Imaging Spectrora- ating in the CS approach require better guidance diometer (MODIS), present specially suitable spectral in plant identification (Morrison 2016, Knapp 2019, and spatial resolutions (13 bands, and pixel size from Bonnet et al 2020) and reported location of phenolo- 250 m to 1000 m) and a long time series (from 2000 to gically active regions of interest (ROIs) (McDonough present) to detect spatial variation during the pheno- MacKenzie et al 2017). They also need protocols logical life cycle. Indeed, MODIS distributes a Global to improve CS skills (MacPhail and Colla 2020, Land Cover dynamics product (Ganguly et al 2010) 2 Environ. Res. Lett. 17 (2022) 085007 C Domingo-Marimon et al that provides 8 d information related to land surface and the RS, creating this interoperable approach, the phenology used in many phenology studies over large scientific community also ensures the availability of geographic regions (Zhang et al 2003, Ahl et al 2006, high-quality research data to be correlated with RS Fisher and Mustard 2007, Zeng et al 2020). How- data, guarantees value for climate impact studies, sup- ever, in many areas of the Earth where fragmented ports public participation in science and enhances and heterogeneous landscape/vegetation is predom- appreciation and engagement in environmental stew- inant, MODIS spatial resolution (250 m at best) is ardship (Larson et al 2020). still too coarse (Nagai et al 2016). Alternatively, Land- Therefore, the aim of this study is to present sat legacy sensors present better spectral and spatial an approach to direct in-situ traditional phenology resolution of eight bands and 30 m, respectively, and volunteers to areas of interest where phenology it can an image archive spanning almost five decades. How- be well characterized by high resolution RS data as ever, Landsat 16 d temporal resolution imagery may well. In this case study, we overcome the crucial limit- be hampered by clouds lowering its ability to generate ing factors of both techniques by (a) generating a pro- a regular and dense time series to detect rapid pheno- tocol that harmonizes CS and RS phenological obser- logical changes such as blossoming, which may occur vations, while (b) evaluating its application regionally rapidly within a 16 d window. to finally, (c) qualitatively evaluate its results. The recent Sentinel-2 satellite European Spatial Agency constellation, (Sentinel-2A and Sentinel-2B) 2. Methodology workflow design and presents some interesting outstanding characterist- application ics compared to its predecessors. Its high spatial (up to 10 m), high spectral (13 bands from visible and A methodological approach (figure 1) combining RS, near-infrared to short wave infrared) and high tem- data and CS was co-designed between scientists and poral (almost only 5 d at the maximum) resolutions CS experts on phenology, and regionally implemen- make it an optimal choice for phenology monitoring ted and evaluated in the study area. (European Commission 2018). This fact was recently To start, a list of species presenting interesting acknowledged by the Copernicus Land Monitoring phenological patterns that could be detected with RS Service (CLMS) by launching a High Resolution data is defined though a co-design session with cit- Vegetation Phenology and Productivity product (HR- izens devoted to monitoring phenology (figure 1(A)). VPP) (Copernicus Land Monitoring Service 2021). Its For each species, specific phenophases are detailed aim is to assess climatic and anthropogenic impacts (figure 2). In parallel, areas with a minimum 3600 m2 on ecosystems and increasing support of the imple- (60 × 60 m) presenting homogeneous vegetation mentation of key EU policies such as the Biodiversity cover and stable phenological responses observable Strategy 2030, the Common Agricultural Policy, the from space are identified based on existing vegeta- Sustainable Development Goals as well as other cli- tion cartography and ancillary data and Sentinel-2 mate initiatives. The HR-VPP product is generated at phenometrics, respectively (Domingo-Marimon et al high spatial resolution (10 m) since 2017 compris- 2020) (figure 1(B)). A set of random 1600 areas is then ing four different daily Vegetation Indices (Normal- selected and for each area three phenometrics (green- ized Difference Vegetation Index—NDVI; Fraction of up, maturity and senescence) based on the Sentinel- Photosynthetically Active Radiation; Plant Phenology 2 NDVI 2019 curve are computed (figure 1(C)). Index and Leaf Area Index), 10 d seasonal trajectories Now that the Copernicus HR-VPP product is avail- and Vegetation Phenological and Productivity (VPP) able, these metrics can be easily extracted and so parameters of yearly metrics for up to two growing it is recommended. Phenometrics are phenological seasons (start or end of the growing season, season phases describing specific stages on the seasonal length, seasonal productivity, etc). growth curve. Green-up is defined as the date (day With this new CLMS product, free and open of year—DOY) on which the slope of the NDVI accessible data through WEkEO Data and Inform- curve increases by 20% in reference to the minimum. ation Access Service (WEkEO 2021), the interact- Maturity is defined as the DOY when the NDVI curve ive portal and distributed public cloud infrastruc- reaches the maximum. Senescence is defined as the ture makes data more available to different users. DOY when the NDVI curve decreases by 80% in ref- Unfortunately, neither the processing environment erence to the minimum. From this dataset, homogen- nor access to HR-VPP products through WEkEO eous vegetation cover areas with a clear and detectable are manageable for the amateur phenology com- phenological pattern (according to clearly differenti- munity who might be interested in considering RS ated phenophases) that are potentially easier to be dif- derived data to help monitor phenology. This lim- ferentiated in the field by citizens are selected as ROIs iting factor reinforces the need for a win–win har- (Domingo-Marimon et al 2020) (figure 1(D)). monized process between CS and RS phenology. By ROIs are used to guide and motivate citizen sci- linking the strengths of both methodologies, the CS entists to collect in-situ phenological observations at 3 Environ. Res. Lett. 17 (2022) 085007 C Domingo-Marimon et al acceptable, then the RS phenometrics can be con- sidered useful to extrapolate phenological observa- tions for all ROIs, dramatically increasing the area ini- tially covered only by CS observations. Workflow in figure 1 can be repeated several times per year coin- ciding with the annual phenological changes resulting in quality assessed RS maps representing the spatial distribution of phenological parameters of individual species or habitats. This workflow was applied in Catalonia, a Medi- terranean region covering 32 000 km2 located on the north–east of the Iberian Peninsula (SW Europe) for two main reasons. First, this area has a wide altitud- inal range (from 0 to more than 3000 m) conferring a heterogeneous landscape with high biodiversity, composed of agriculture land (herbaceous and orch- ards) mixed with natural and semi-natural vegeta- tion areas, including coniferous, sclerophyllous, and deciduous forest. Second, and most importantly, Catalonia Figure 1. Workflow of the developed new approach stands out for having a wide and strong com- combining RS and developed tools to facilitate and foster munity of CS observers thanks to the FenoCat phenology CS. network-Phenological Network of Catalonia (Servei Meteorològic de Catalunya 2020) and the Rit- meNatura CS Observatory (Ritme Natura Observat- ory 2015) initiatives, thus facilitating the engagement of volunteers to participate and evaluate the proposed methodology. Thus, due to these aspects, the study area is a unique spot for evaluating an interoperable approach between CS and RS phenology monitoring. In this study, ROIs selection was carried out com- bining the Map of Habitats (Generalitat de Catalunya 2018), the Land Parcel Identification System of Catalunya (Departament d’Agricultura, Ramaderia, Pesca i Alimentació, LPIS 2019), the Land use and Land cover Map of Catalunya (Ibáñez and Burriel 2010), the 4th National Forestry Inventory (Ministerio para la Transición Ecológica y el Reto Demográfico 2008) and the Forest Map of Spain (1:50 000) (Ministerio de la Transición Ecológica y el Reto demográfico 2006). The resulting dataset was stored into a data visu- alization portal developed using MiraMon Map Browser technology that fulfils the most common Figure 2. App phenophase selection options provide Open Geospatial Consortium (OGC) standards specific phenological information per each species. (A) Phenophases observed for Rapeseed, (B) phenophases (FenoTwin-PhenoTandem 2020). observed for Sweet Chestnut. 3. Results, outcomes and discussion The methodology presented here is extensible to any these particular areas through a data collection app location with similar vegetation phenological beha- (figure 1(E)). viour. By repeating this approach in subsequent years, Finally, once the phenological season ends, the it is possible to assess the impact of climate change on CS date of observations of a specific event of a given each species over time. Also, under the assumptions ROI can then be compared with the same events that the methodology has a stable degree of quality extracted by RS (figure 1(F)). The number of times and that species distribution has not changed signific- the two observations coincide will give an indication antly in different periods, this methodology is applic- of the quality of the RS product (uncertainty) that able to archived high resolution RS data, thus extend- can be used for product evaluation. If the quality is ing the time series. 4 Environ. Res. Lett. 17 (2022) 085007 C Domingo-Marimon et al 3.1. Map of regions of interest (ROIs) CS observations. Since the observations are done in As a result of the co-design session, a list of 41 spe- homogeneous ROIs, the species or habitat identifica- cies presenting discernible phenological patterns that tion is automated by position. Unlike other success- can be both detected in-situ and with RS data was ful applications such as iNaturalist (the global CS obtained. Combining this list with a map of homo- platform for biodiversity monitoring) listing only the geneous and pure habitat or species (60 × 60 m areas) most common phenophases, the data collection app resulted in ROIs dataset indicating where in-situ developed in this study offers the unique phenological observations were comparable with RS observations phases relevant to the species observed, thus minim- and when and which phenophase was detectable ising likely identification errors made by observers and active for a given species and when this event and reducing the uncertainty of CS data. The data took place in the near past (Domingo-Marimon et al collection app is displayed as a simple form based 2020). on six steps: (a) uploading a picture, (b) selecting a species from a list (automatically detected from 3.2. Visualization portal for CS the ROI layer), (c) selecting a particular pheno- A single interface was developed to integrate all res- phase, (d) reporting an observation location (auto ulting products to facilitate CS accessibility to data filled if the device GPS is activated), (e) providing (Masó 2022). The Visualization Portal for CS obser- an observation date (auto filled with ‘today’) and, vations resulted in a user-friendly interface showing (f) advanced options for phenophase species identi- a ROI distribution map enabling citizens to find spe- fication (figure 2). Each step includes illustrative pic- cific areas for a homogeneous species closer to their tures and understandable semantics as well as user location (FenoTwin-PhenoTandem 2020). In addi- field protocols describing the species and pheno- tion to this functionality, it also provided an integ- phases available to guide citizens during identification rated digital compass supporting users’ orientation in field campaigns. In the 2020 autumn campaign, and positioning, when the global positioning system more than 200 in-situ observations for selected spe- (GPS) of the device (smartphone or PC) is activated cies in specific ROIs were collected. (Domingo-Marimon et al 2020). Therefore, users are guided to ROIs rather than those typically monitored 3.4. Quality of CS observations—user feedback (e.g. areas close to CS homes, close to pathways or In addition to the iNaturalist independent valida- easy accessible areas), endowing their observations tion system based on its community of contribut- with a new dimension. The map browser also shows a ing experts, a third quality component was included high-resolution phenometrics maps displaying when into the data capture system: a feedback mechanism. a particular phenological phase occurred in a pre- NiMMbus (Masó and Zabala 2019) is a widget to vious year for a given species. This feature allows provide user feedback items following the OGC geo- users to generate custom data visualization, to per- spatial user feedback standard such as, comments, form some direct analytical operations and can be ratings, discovered issues, quality reports among oth- extended with an animation functionality provid- ers. Citizens participating in data collection cam- ing temporal data profiles (e.g. NDVI profiles) that paigns are able to comment on any of the geospatial can be correlated with CS observations. Also, direct features in the system. This includes a ROI, phe- clicking on a particular ROI returns the geographic nometrics products, another citizen observation or coordinates, the dominant species and a link redir- even suggest new relevant ROIs that have not been ecting users to the data collection website applica- initially included by the RS algorithms (figure 1(G)). tion. In addition, the visualization of the map browser All feedback is made available to everyone and can be layer-enriched by activating any type of envir- linked to a relevant geospatial feature. onmental data, such as topographic maps or digital elevation models. The resulting in-situ observations 3.5. CS engaging campaigns collected by citizens are also included in the portal Since phenology is a topic directly related to nature, and are later used to evaluate Sentinel-2 derived phe- the profile of a citizen who can potentially become nometric products. engaged might be someone who is environment- ally concerned. That is the reason the monitor- 3.3. Data collection app ing protocol was co-designed with local nature, A freely available Data Capture Web App to facilitate hiking and meteorology associations. By making field data collection from volunteers was developed them directly involved in the research process (by (Masó 2020). This tool transforms in-situ data acquis- deciding species and phenophases to be observed), ition into an easy and attractive task. The app’s their participation in the engaging campaigns was user-friendly online interface is fully connected to ensured. Three sessions took place at the beginning, the Visualization Portal, giving the user a seamless middle and end of the implementation process. experience in the harmonization of RS data and Two information campaigns to recruit volunteers 5 Environ. Res. Lett. 17 (2022) 085007 C Domingo-Marimon et al Figure 3. Top maps: Sentinel-2 derived phenometrics of green-up, maturity and senescence for a region with F. sylvatica (dashed line) and a region with Q. pubescens (black line). Bottom graph: area (in ha) per phenophase and DOY of F. sylvatica (dashed line) and Q. pubescens (black line). These distributions are essential to guide volunteers to precise locations at the precise time of the year. were defined in 2020, one in spring and another 3.6. Harmonization and validation of CS and RS in autumn. The campaigns included nature associ- phenology data ations newsletters, hikers’ magazine, meteorological Results of the analysis of Sentinel-2 phenometrics associations, and related channels on social networks for 2019 (figure 3) showed clear and different phen- (mainly Twitter, TV and digital newsletters). Around ological patterns for several species, especially for 50 000 citizens received direct information through deciduous species, such as Fagus sylvatica or Quercus these media channels. The call also included an initial pubescens. Results showed when and where each spe- press release published in several specialized media cies were leafing out (green-up), when it was at the and blog sites, and regular notices explaining the peak of its foliage (maturity) and when it started to progress of phenological events to guide citizens to lose leaves (senescence), confirming Sentinel-2 phe- phenological attractive spots pre-detected through nometrics as of good quality. The information associ- RS. Campaigns addressed to specific individuals were ated with these distributions is essential to guide cit- more effective than mass media approaches. The izens to precise locations at the precise time of year objective of the campaigns was clearly described to to be able to collect in-situ data. By assuming the increase the chances of success. A short, clear, and quality of the computed phenometrics as acceptable, concise monitoring protocol (including relevant spe- the number of ROIs was extrapolated to more than cies fact sheets) was also distributed to ensure good 5000. Although more in-situ observations are needed quality of data generated. A web page and contact to perform a complete and solid integration of RS and email were set up to support volunteers in the field. in-situ data, a first qualitative correspondence analysis Unfortunately, the spring campaign was completely between RS observations and CS observations yiel- affected by the COVID-19 pandemic. Citizens were ded a good visual agreement between both types of confined with no possibility of taking observations. data (figure 4). In case of several observations for a Although the autumn campaign was still affected by given species and therefore, several DOY indicating COVID-19 restrictions, it recruited 12 direct volun- observed phenophases, the DOY was used to analyse teers and 244 occasional observers. within species phenophase variability. 6 Environ. Res. Lett. 17 (2022) 085007 C Domingo-Marimon et al collected indicated that a user-friendly data acquisi- tion interface is desirable to make observations easier for citizen scientists. The methodological strategy based on Sentinel- 2 and the newly developed tools were effective and showed that a harmonized approach between phen- ological high resolution optical RS and CS observa- tions is operative, functional and practicable. It was also shown that satellite RS observations, especially those presenting high spatial and temporal resolutions such as Sentinel-2, have the capability to spatiotemporally characterize vegetation pheno- logy in selected ROIs. This overcomes traditional lim- itations of citizens’ observations, such as reduced extension or temporal continuity. The presented approach uses a combination of existing cartography and Sentinel-2 data to spatiotemporally spot phen- ological ROIs and their phenophases to guide CS volunteers. In return, the improvement in quantity and quality of data collected by CS ensures its usefulness in scientific RS studies, as the validation data increases significantly. Such information is displayed through a newly developed Phenometrics Visualization Portal. Due to COVID-19 pandemics, a limited CS obser- vations dataset was collected. Although this allowed for a visual comparison of both RS data and CS Figure 4. Qualitative correspondence between the observations yielding a visual consistent phenolo- phenological observations extracted from RS NDVI time gical curves, a more extended CS dataset is needed to series images (green dots) and interpolated trend (green evaluate RS phenology products quantitatively. Thus, line) and the in-situ observations collected during autumn 2020 (coloured triangles, their vertical position is future work will be developed to quantitatively eval- arbitrary). uate the new methodology. Temporal profiles of RS vegetation indices related to greening, maturity and senescence were coher- 4. Conclusions ent with the same events observed at ground phen- ologically active ROIs. As a result of this methodo- Phenology monitoring has become one of the main logy, around 5000 new areas presenting phenological indicators of climate change impacts on vegetation. Is changes were identified, significantly increasing ROIs essential to keep track of in-situ vegetation changes, to be visited by CS, as well as the number of observed especially by citizen scientists around the world. species. Accuracy of greening-related events from RS This paper describes a co-designed observation pro- is higher, although less engaging for citizens who are tocol between scientists and potential volunteers to usually more attracted to blossoming related events. improve citizen engagement and the quality of data This has turned into future work, since improving collected, by providing simple and clear instructions the detection accuracy of blossoming from space and tools. This protocol can be applied to any region, would integrate RS more into the citizen’s perspective biome or habitat several times per year and can be and classical phenology. In summary, the implemen- adjusted annually according to new ROIs, species of ted methodology was successfully applied, increas- interest or changes to phenology events. This protocol ing the quality and number of phenological obser- provides an innovative approach, making CS obser- vations recorded, establishing RS as the new travel vations interoperable with RS products and confirms companion of CS chasing phenology around the RS as a promising partnership for public participation world. in environmental research activities. A lesson learned from this situation is that targeted dissemination to Data availability statement interest groups was more effective than mass media campaigns. Therefore, relying on individual commit- The data that support the findings of this study ment rather than an organizational or institutional are openly available at the following URL/DOI: one is recommended. In addition, the feedback www.inaturalist.org/projects/ritmenatura. 7 Environ. Res. Lett. 17 (2022) 085007 C Domingo-Marimon et al Funding https://land.copernicus.eu/pan-european/biophysical- parameters/high-resolution-vegetation-phenology-and- productivity) This work has been partially funded by the European Crimmins T, Denny E, Posthumus E, Rosemartin A, Croll R, Spatial Agency (ESA) through the CSEOL Pro- Montano M and Panci H 2022 Science and management gramme, the H2020 Ground Truth 2.0, WeObserve advancements made possible by the USA National and Cos4Cloud projects, which received funding Phenology Network’s Nature’s Notebook platform (available at: www.essoar.org/doi/10.1002/essoar.10509152.2) from the European Union’s Horizon 2020 Research de Beurs K M and Henebry G M 2004 Land surface phenology, and Innovation Programme under Grant Agreement climatic variation, and institutional change: analyzing Nos. 689744, 776740 and 863463 respectively and agricultural land cover change in Kazakhstan Remote Sens. by the FenoTwin Project (FCT-20-16181), funded Environ. 89 497–509 Departament d’Agricultura, Ramaderia, Pesca i Alimentació 2019 by FECYT (Spanish Foundation for Science and Land parcel identification system of Catalunya (LPIS) Technology). Dickinson J L, Zuckerberg B and Bonter D N 2010 Citizen science as an ecological research tool: challenges and benefits Annu. Rev. Ecol. Evol. Syst. 41 149–72 Acknowledgments Domingo-Marimon C, Masó J, Cristóbal J, Batalla M and Ninyerola M 2020 Remote sensing as a driving tool for The authors are grateful to three anonymous review- citizen science phenology monitoring campaigns 8th Int. Conf. on Remote Sensing and Geoinformation of the ers for their useful comments that improved this Environment (Rscy2020) vol 11524 (SPIE) pp 596–605 paper significantly. The authors are also grate- (available at: www.spiedigitallibrary.org/conference- ful for all the citizen observers that have con- proceedings-of-spie/11524/115241N/Remote-sensing-as-a- tributed phenology records to Natusfera and iNat- driving-tool-for-Citizen-Science-phenology/10.1117/ 12.2570842.full) uralist under the RitmeNatura project over the Duchemin B, Guyon D and Lagouarde J P 1999 Potential and years. limits of NOAA-AVHRR temporal composite data for phenology and water stress monitoring of temperate forest ORCID iDs ecosystems Int. J. Remote Sens. 20 895–917 European Commission 2018 Annex to the Commission Implementing Decision on the Adoption of the Work Cristina Domingo-Marimon  Programme 2018 and on the Financing of the Copernicus https://orcid.org/0000-0001-6822-8704 Programme (Brussels: European Commission) (available at: Joan Masó  https://orcid.org/0000-0002-2983- http://ec.europa.eu/transparency/regdoc/rep/3/2018/EN/C- 2018-2-F1-EN-ANNEX-1-PART-1.PDF) 4629 Evans S G, Small E E and Larson K M 2014 Comparison of Ester Prat  https://orcid.org/0000-0001-9475-4070 vegetation phenology in the western USA determined from Alaitz Zabala  https://orcid.org/0000-0002-3931- reflected GPS microwave signals and NDVI Int. J. Remote 4221 Sens. 35 2996–3017 Feldman R E, Žemaitė I and Miller-Rushing A J 2018 How Ivette Serral  https://orcid.org/0000-0002-7651- training citizen scientists affects the accuracy and precision 656X of phenological data Int. J. Biometeorol. 62 1421–35 Meritxell Batalla  https://orcid.org/0000-0002- FenoTwin-PhenoTandem 2020 FenoTwin map browser (available 6378-9681 at: www.ritmenatura.cat/projects/fenoTwinBrowser Produccio/index.htm) Miquel Ninyerola  https://orcid.org/0000-0002- Fisher J I and Mustard J F 2007 Cross-scalar satellite phenology 1101-0453 from ground, Landsat, and MODIS data Remote Sens. Jordi Cristóbal  https://orcid.org/0000-0001-6244- Environ. 109 261–73 4289 Forrester T D, Baker M, Costello R, Kays R, Parsons A W and McShea W J 2017 Creating advocates for mammal conservation through citizen science Biol. Conserv. References 208 98–105 Fraisl D et al 2020 Mapping citizen science contributions to the Ahl D E, Gower S T, Burrows S N, Shabanov N V, Myneni R B and UN sustainable development goals Sustain. Sci. 15 1735–51 Knyazikhin Y 2006 Monitoring spring canopy phenology of Ganguly S, Friedl M A, Tan B, Zhang X and Verma M 2010 Land a deciduous broadleaf forest using MODIS Remote Sens. surface phenology from MODIS: characterization of the Environ. 104 88–95 collection 5 global land cover dynamics product Remote Bonnet P et al 2020 How citizen scientists contribute to monitor Sens. Environ. 114 1805–16 protected areas thanks to automatic plant identification García M B, Silva J L, Tejero P and Pardo I 2021 Detecting tools Ecol. Sol. Evid. 1 e12023 early-warning signals of concern in plant populations with a Bonney R, Shirk J L, Phillips T B, Wiggins A, Ballard H L, citizen science network. Are threatened and other priority Miller-Rushing A J and Parrish J K 2014 Next steps for species for conservation performing worse? J. Appl. Ecol. citizen science Science 343 1436–7 58 1388–98 Burgess H K, DeBey L B, Froehlich H E, Schmidt N, Theobald E J, Generalitat de Catalunya 2018 Map of Habitats of Catalunya v.2 Ettinger A K, HilleRisLambers J, Tewksbury J and Parrish J K (available at: http://territori.gencat.cat/ca/01_departament/ 2017 The science of citizen science: exploring barriers to use 12_cartografia_i_toponimia/bases_cartografiques/medi_ as a primary research tool Biol. Conserv. 208 113–20 ambient_i_sostenibilitat/bases_miramon/territori/29_ Chandler M et al 2017 Contribution of citizen science towards habitats_1_5000_perfulls/) international biodiversity monitoring Biol. Conserv. Guo H 2014 Special section guest editorial: Earth observation for 213 280–94 global environmental change J. Appl. Remote Sens. 8 084501 Copernicus Land Monitoring Service 2021 High resolution Ibáñez J J and Burriel J A 2010 Mapa de cubiertas del suelo de vegetation phenology and productivity (available at: Cataluña: características de la tercera edición y relación con 8 Environ. Res. Lett. 17 (2022) 085007 C Domingo-Marimon et al SIOSE Tecnologías de la Información Geográfica: La normalized-difference vegetation index for phenology Información Geográfica al servicio de los ciudadanos ed J detection Agric. For. Meteorol. 150 152–60 Ojeda, M F Pita and I Vallejo (Sevilla: Secretariado de Peñuelas J, Rutishauser T and Filella I 2009 Phenology feedbacks Publicaciones de la Universidad de Sevilla) on climate change Science 324 887–8 pp 179–98 Phillimore A B, Proios K, O’Mahony N, Bernard R, Lord A M, Isaacson B, Serbin S and Townsend P 2012 Detection of relative Atkinson S and Smithers R J 2013 Inferring local processes differences in phenology of forest species using Landsat and from macro-scale phenological pattern: a comparison of two MODIS Landsc Ecol. 27 529–43 methods J. Ecol. 101 774–83 Kirchhoff C, Callaghan C T, Keith D A, Indiarto D, Taseski G, Pucino N, Kennedy D M, Carvalho R C, Allan B and Ooi M K J, Le Breton T D, Mesaglio T, Kingsford R T and Ierodiaconou D 2021 Citizen science for monitoring Cornwell W K 2021 Rapidly mapping fire effects on seasonal-scale beach erosion and behaviour with aerial biodiversity at a large-scale using citizen science Sci. Total drones Sci. Rep. 11 3935 Environ. 755 142348 Richardson A D, Keenan T F, Migliavacca M, Ryu Y, Sonnentag O Knapp S 2019 Are humans really blind to plants? Plants People and Toomey M 2013 Climate change, phenology, and Planet 1 164–8 phenological control of vegetation feedbacks to the climate Larson L R, Cooper C B, Futch S, Singh D, Shipley N J, Dale K, system Agric. For. Meteorol. 169 156–73 LeBaron G S and Takekawa J Y 2020 The diverse motivations Ritme Natura Observatory 2015 Ritme Natura Observatory of citizen scientists: does conservation emphasis grow as (available at: http://ritmenatura.cat/index-eng.htm) volunteer participation progresses? Biol. Conserv. Schuetz J G, Mills K E, Allyn A J, Stamieszkin K, Bris A L and 242 108428 Pershing A J 2019 Complex patterns of temperature MacPhail V J and Colla S R 2020 Power of the people: a review of sensitivity, not ecological traits, dictate diverse species citizen science programs for conservation Biol. Conserv. responses to climate change Ecography 42 111–24 249 108739 Scranton K and Amarasekare P 2017 Predicting phenological Masó J 2020 Phenology data capture web app (available at: shifts in a changing climate Proc. Natl Acad. Sci. www.ritmenatura.cat/captura/index.htm?Language=eng) 114 13212–7 Masó J 2022 MiraMonMapBrowser (Grumets Research Group) Servei Meteorològic de Catalunya 2020 FENOCAT, la xarxa (available at: https://github.com/grumets/MiraMon fenològica de Catalunya Servei Meteorològic de Catalunya MapBrowser) (available at: www.meteo.cat/wpweb/climatologia/el-clima- Masó J and Zabala A 2019 NiMMBus: geospatial user feedback ara/fenologia/) implementation (available at: https://github.com/ Shirk J et al 2012 Public participation in scientific research: a joanma747/nimmbus) framework for deliberate design Ecol. Soc. 17 29 McDonough MacKenzie C, Murray G, Primack R and Thuiller W 2004 Patterns and uncertainties of species’ range shifts Weihrauch D 2017 Lessons from citizen science: assessing under climate change Glob. Change Biol. 10 2020–7 volunteer-collected plant phenology data with Mountain WEkEO 2021 WEkEO platform (available at: www.wekeo. Watch Biol. Conserv. 208 121–6 eu/news/clms-releases-hr-vpp-product-to-assess- McKinley D C et al 2017 Citizen science can improve conservation ecosystems-and-biodiversity) science, natural resource management, and environmental White M A et al 2009 Intercomparison, interpretation, and protection Biol. Conserv. 208 15–28 assessment of spring phenology in North America estimated Ministerio de la Transición Ecológica y el Reto demográfico 2006 from remote sensing for 1982–2006 Glob. Change Biol. Forest map of Spain (available at: www.miteco.gob.es/es/ 15 2335–59 biodiversidad/servicios/banco-datos-naturaleza/ White M A, Hoffman F, Hargrove W W and Nemani R R 2005 A informacion-disponible/mfe50.aspx) global framework for monitoring phenological responses to Ministerio para la Transición Ecológica y el Reto Demográfico climate change Geophys. Res. Lett. 32 1–4 2008 Cuarto Inventario Forestal Nacional (IFN4) (available White M A, Thornton P E and Running S W 1997 A continental at: www.miteco.gob.es/es/biodiversidad/temas/inventarios- phenology model for monitoring vegetation responses to nacionales/inventario-forestal-nacional/cuarto_ interannual climatic variability Glob. Biogeochem. Cycles inventario.aspx) 11 217–34 Morrison L W 2016 Observer error in vegetation surveys: a review Zeng L, Wardlow B D, Xiang D, Hu S and Li D 2020 A review of J. Plant Ecol. 9 367–79 vegetation phenological metrics extraction using Nagai S, Nasahara K N, Inoue T, Saitoh T M and Suzuki R 2016 time-series, multispectral satellite data Remote Sens. Environ. Review: advances in situ and satellite phenological 237 111511 observations in Japan Int J Biometeorol 60 615–27 Zhang X, Friedl M A, Schaaf C B, Strahler A H, Hodges J C F, Nagai S, Nasahara K N, Muraoka H, Akiyama T and Tsuchida S Gao F, Reed B C and Huete A R 2003 Monitoring vegetation 2010 Field experiments to test the use of the phenology using MODIS Remote Sens. Environ. 84 471–5 9