INTRODUCTION

Breeding is an energetically demanding event in the life history of birds that can directly influ ence parental fitness and population persistence (^{Berl et al., 2014}; ^{Frei et al., 2015}). Detailed de scriptions on breeding biology provide natural history data that describe patterns of variation and can be used to address questions of evolu tionary, ecological, or conservation significance. Knowledge of basic aspects of breeding biology, such as the breeding season, nest characteristics, clutch size, and nest success is well document ed for most Nearctic and Holarctic birds (^{Xiao et al., 2017}) and has helped to understand the influence of phylogenetic history of species on such reproductive traits, the influence of selec tive pressures on reproductive strategies, or to implement sound management and conserva tion actions for these birds and their habitats (^{Martin, 1995}; Martin & Clobert, 1996; ^{Keller, 2014}; ^{Böhning-Gaese et al., 2000}; ^{Hudson et al., 2017}). Unfortunately, the information avail able about basic aspects of breeding biology is still insufficient for Neotropical birds, being known for only 19% of these species (Xiao et al., 2017). Several neotropical countries, including Venezuela, Argentina, Brazil, Paraguay, Ecuador, and Colombia, have made great progress in pro viding the first available information about the breeding biology of hundreds of bird species, but this information is mostly anecdotal and has been collected on a short-term basis (^{Fierro-Calderón et al., 2021}). Notable exceptions of sys tematic long-term studies are found in Venezuela and Argentina, which have allowed large-scale comparisons of some reproductive traits (i.e., territory size, incubation period, egg size varia tion, clutch size, brood size, and nestling growth rates) between north-temperate or tropical birds and south temperate birds (Martin, 2008; Martin et al., 2011; ^{Llambías et al., 2015}; Martin et al., 2018); or Panama and Brazil, from which studies of nest success have increased our understanding of predation and forest fragmentation pressures in tropical forests (^{Robinson et al., 2000}; ^{Libsch et al., 2008}; ^{Borges & Marini, 2010}). Information about the breeding biology of Peruvian birds is also at the frontier of exploration. Most pub lished data comes from first nest descriptions, anecdotal observations in general avifaunal in ventories, and monitoring of nesting and incu bation behavior in remote places (such as guano islands, desert scrubs, high-Andean region, hu mid eastern montane forest, and lowland rain forest) (^{Franke, 2017}), while only a few studies have focused on the breeding biology of birds in habiting urban areas (^{Gonzalez, 1998}; Gonzalez et al., 1999; Gonzalez, 2004; ^{Ortiz, 2012}; Ortiz, 2013; ^{Tavera, 2011}; ^{Rivas et al., 2013}; ^{Amaro & Goyoneche, 2017}; ^{Angulo & Moran, 2019}; ^{Arenas et al., 2020}; ^{Díaz et al., 2022}).

Methods that explore the breeding biology of birds include systematic search (i.e., nest search ing, nest monitoring, or spot mapping), and the capture and marking of birds; however, these methods often take a long time for researchers to compile a significant amount of data (^{Alegria, 2018}). Furthermore, these approaches can be very labor intensive and costly, so most studies are limited to a small geographic area, a small proportion of the population, or a small group of target species (^{Ralph et al., 1993}). Large data bases are needed to gain a comprehensive under standing of when, where, and how Neotropical birds breed. The use of citizen science databases as a potential new method to examine the breed ing biology of birds might overcome the limita tions mentioned above (^{Fierro-Calderón et al., 2021}). The main purpose of citizen science por tals is to report the occurrence/abundance of bio diversity, but also, optionally, additional details from the sightseeing can be provided, including the organism’s age, sex, behavior, observation ef fort, etc. Additionally, the advances in computing and communication technology now make it pos sible for observers and researchers to upload me dia files (i.e., photography, videos, and sounds) on these portals, resulting from their observa tions in the field from anywhere and at any time (^{Pocock et al., 2018}). Through these widespread networks of observers, millions of observational bird reports accumulate across multiple sites all over the world throughout the year, providing an unprecedented trove of information. This infor mation might include a considerable number of breeding records throughout a species’ annual cycle, however, no efforts related to an in-depth examination of bird records retrieved from citi zen science portals have yet been published ac cording to a review of the “Bibliography of birds of Peru” (^{Plenge, 2020}).

Currently, eBird (https://ebird.org) and iNat uralist (https://inaturalist.org) are two impor tant worldwide citizen portals used by Peruvian birdwatchers. These portals host tens to hun dreds of thousands of records across the coun try (eBird, 2022; iNaturalist, 2022), but possess different data models and different data quality control strategies. eBird is considered a “semi-structured” database (i.e., when the design of the database allows extracting of information on sampling effort or completeness) managed by the Cornell Lab of Ornithology (^{Sullivan et al., 2009}). Data quality in this platform is controlled based on accepted species distributions and their estimated counts and suspected unusual records are then reviewed by regional expert reviewers (^{Wood et al., 2011}). On the other hand, iNatural ist is an “unstructured” database (i.e., when the database lacks relevant information on sampling effort) managed by the California Academy of Sciences that allows users to submit single ob servations of multiple taxa even without previ ous knowledge about specific taxonomic groups (^{Nugent, 2018}). Data quality in this platform is controlled based on a community identification process but also facilitated by a machine learn ing algorithm that makes suggestions on species identification to the participant (^{Van Horn et al., 2018}).

Neotropical cities are growing rapidly, ex panding over native habitats, and generally conserving only small green areas (^{de Camargo Barbosa, 2021}). A paradigmatic case is the Lima Metropolitan Area (LMA) (Fig. 1), which is con sidered the most populous and the largest met ropolitan area in Peru because it hosts over 30% of the total national population (10.7 million inhabitants) settled in only 0.2% of the nation al territory (2811.65 km²) (Instituto Nacional de Estadística e Informática, 2015; Instituto Nacional de Estadística e Informática, 2020). The LMA has a vegetation coverage, made up of hills, wetlands, valleys, and green areas, which only represent 14.4% of its total area (405.29 km²,) (Lima Cómo Vamos, 2014). Despite such a small area, these spaces can still harbor important bird diversity including native, exotic, migrant, and rare species (^{Nolazco, 2012}). Given the limited information on the breeding biology of Peruvian urban birds, even opportunistic data compiled from citizen science projects merits publication. The aim of this study was threefold: (1) to de termine the geographical distribution of breed ing records and patterns of breeding activity, (2) to examine the timing of breeding activity, and (3) to describe patterns of nesting behavior (i.e., nesting habitat use, host plant preference, and clutch size) based on bird observational records retrieved from two citizen science portals, eBird and iNaturalist, from the LMA during the 2000-2020 period (Figs. 2, 3). We also discuss the ben efits and challenges of using citizen science data to study Peruvian birds’ breeding biology.

Fig. 1 Map of the Lima Metropolitan Area and the density of breeding records by district retrieved from eBird’s and iNaturalist’s citizen science data from January 2000 to December 2020. 

Fig. 2 Number of breeding records per species, classified by their activity, retrieved from eBird and iNaturalist’s citizen science data from January 2000 to December 2020. 

Fig. 3 Number of nesting records by habitat type retrieved from eBird and iNaturalist’s citizen science data from January 2000 to December 2020. Occurrence status in LMA: N = native species, I = introduced species. 

METHODS

Study area. The Lima Metropolitan Area (LMA) is in the central and western zone of the South American continent (longitude 77°W and latitude 12°S), in Peru, in the department of Lima. It is an area formed by the conurbation of the Peruvian provinces of Lima and Callao that extends on a large alluvial plain formed by the valleys of the Chillón, Rímac, and Lurin rivers (^{Rojas et al., 2021}). The LMA is considered to have a sub tropical desert climate with almost no rainfall throughout the year (average annual precipita tion is 10 mm), a temperature range of 14°C to 27°C (average annual temperature is 19°C), and relative humidity of 70% to 100% (average an nual humidity is 80%) (SENAMHI, 2016). The dry-warm season occurs from December to April, and the humid-cold season from June to October, with May and November as transition months (^{Reátegui-Romero et al., 2018}). Dominant habi tat substrates within the study area can be classi fied into three major groups following ^{Tello-León (2021}): green space, referred to as areas covered with vegetation, including parks, gardens, street trees, and coastal hills; blue space, referred to as waterbodies or watercourses in the city, includ ing artificial reservoirs, river valleys, wetlands, and beaches; and gray spaces, referred to man-made structures formed by paved areas with a civic function, including streets, parking areas, buildings, and utility poles (Online Appendix 1).

Datasets used in this study and descrip tion of breeding records. We included photos and videos that evidenced breeding activity col lected from 1st January 2000 to 31st December 2020, across 50 municipal districts of the Lima Metropolitan Area (including 43 from Lima Province and 7 from Callao Province), from eBird and iNaturalist. We classified breeding re cords into nine categories: courtship or copula tion (C), adult carrying nesting material (CN), occupied nest (ON), nest with egg (NE), active nest at unknown stage (AN), nestling (N), feed ing young (FY), carrying food for young (CF), and recently fledged young (FL). We considered evidence of breeding for non-passerine and pas serine birds, including waterfowl, shorebirds, waterbirds, and landbirds. We did not consider evidence of adult birds of certain species spotted on burrows, cavities, or holes, including parrots, wrens, owls, and swallows, because they use it not only for nesting but also for roosting. We also excluded cases of potential hybrids. Descriptions of breeding events included, when possible, in formation on mating behavior, clutch size, nest location, the identity of the plant supporting the nest, and parental and young behavior (Online Appendix 2). Age terminology for young birds provided in descriptions followed the consensus definitions proposed by ^{Wood (1946}), includ ing “nestling”, a bird within and not ready to leave the nest; “fledgling”, a bird that has grown enough to acquire its initial flight feathers and is preparing to leave the nest and survive but still being cared for by its parents; “juvenile”, a bird in its first plumage of non-downy feathers (juvenal plumage) that has left the nest and is entirely independent; and “immature”, a bird in any non-adult plumage, including (but not lim ited to) the juvenal plumage. The juvenal plum age is very short-lived in most passerines com pared to non-passerine birds (^{Pittaway, 2000}) and certain neotropical passerines can disperse long distances even wearing such plumage (^{Pyle et al., 2015}; ^{Gorleri & Areta, 2021}), hence, cases of “juveniles” and “immature” birds were ex cluded from our collection of breeding records given that the goal of our study was to include the closest reliable records in time and place to the mating event.

Preparation of citizen science datasets. We used eBird basic datasets (data upon request on 10 July 2021; version ebd_Jun-2021) and in cluded data from any type of observational pro tocol (stationary, travelling, incidental, histori cal, or other specialized sampling protocols) (For more details on the eBird methodology, see here: https://support.ebird.org/en/support/home). iNat uralist datasets only included observation reach ing the “research grade” status, i.e., when two-thirds of the identifiers of the iNaturalist com munity confirm the species-level identification for a given observation (For more details on the iNaturalist methodology, see here: https://inatu ralist.org/pages/getting+started). Filtered data status from both citizen science portals does not necessarily ensure that the identification of spe cies are correct, as even knowledgeable users and reviewers can sometimes make mistakes (^{Austen et al., 2018}; ^{Callaghan et al., 2021}; ^{Rocha-López et al., 2021}; ^{Gorleri & Areta, 2021}). Moreover, information about breeding behavior of birds might not necessarily be explicit or correctly ex pressed solely by notes or standardized codes in records from both portals. To account for these biases, we thoroughly scrutinized records with photographs or videos that allowed us to reli ably evidence some aspect of the breeding biol ogy of birds. Furthermore, in some cases, eBird checklists and iNaturalist observations might contain multiple images of different individuals that allow us to evidence more than one type of breeding event. For these cases, we treated im ages showing different types of breeding events as independent records. We also ordered images by date and observer’s name to allow a better tracking of individuals from breeding records, thus minimizing the duplication of data from both portals. Given that the information about the breeding biology of Peruvian birds is poor ly know, limited data available from eBird and iNaturalist projects related to this aspect could also be expected. Hence, breeding data from both portals were combined for descriptive analyses in this study.

A systematic search was also performed on two specialized Facebook groups, named “Aves de Lima”, and “Aves Del Peru”, on 10th December 2021, using the following keywords “reproduc ción” (breeding), “apareamiento” (mating), “copula” (copulation), “nido” (nest), “huevo” (egg), “polluelo” (nestling), “pichon” (nestling), “volantón” (fledgling), to locate records pub lished by users. We only considered records that were not previously published on the eBird and iNaturalist citizen portals. When the informa tion found on one Facebook group was repeated on another, only the first was considered. Then we asked users to submit their breeding records to any of the previously mentioned citizen sci ence portals. An eBird’s group account named “Biologia Reproductiva de Aves Peruanas” was created to incorporate records of users who did not intended to create their own account but pro vided us permission to submit their records with the assurance that the images would not be used for other purposes. In addition, we created an iNaturalist’s traditional project and announced it in Facebook groups to encourage users to sub mit additional unreleased breeding records to such citizen science portal.

Geographic distribution of breeding re cords and breeding activity of species. Since datasets from these citizen science portals did not provide location data at the district level, we intersected observations (based on their lati tude/longitude reported location) with a shape file of Peruvian districts retrieved from Geo GPS Perú (2014) using QGis 2.14 (Quantum GIS Development, 2016). The geographic distribu tion of breeding bird records was examined by plotting a choropleth map showing the frequency of the different types of breeding activity record per district of the LMA. Breeding species and the types of breeding activity found were represent ed in a stacked bar graph showing the frequency of each type of breeding record per species.

Breeding seasonality. The timing of breeding activity for each species was examined by sum ming up all types of breeding records per month to establish periods of breeding activity. Breeding months were then compared to preliminary data for the central coast of Peru (latitudes ~9.5°S to ~14.5°S) to determine coincidences that would corroborate or improve the knowledge of prelimi nary information. Comparisons followed the cri teria proposed by ^{Verea et al. (2009}), in which the timing of breeding is considered as “coincident” when the breeding months matches identically or falls within the breeding period preliminary proposed for a particular species; “improved”, when there was no preliminary information or if, after coinciding with it, additional months of breeding activity were observed; and “noncoinci dent”, when the breeding months did not coin cide in any way with preliminary data. We ana lyzed breeding periods for birds that showed at least four continuous monthly records of breed ing activity. Also, we considered a species able to breed throughout the year for those whose num ber of breeding months was equal to or greater than nine

Patterns of nesting behavior. Description of nesting patterns was based on records of nesting birds, including categories of “occupied nest”, “nest with egg”, “active nest at unknown stage”, and “nestling”. Nesting-habitat use was de scribed by plotting a stacked bar graph showing the type of habitat used (green, blue, and gray spaces) per each species of nesting birds. Host plant preference was described by showing a list of identified plant species that provided support to each species of nesting bird by habitat type. We excluded cases of photos or videos that made it harder to correctly identify plant species, es pecially the ones with poor focus or inadequate zooming on the host plants. Finally, clutch fre quency of each species per habitat was described based solely on the records belonging to the cat egory “nest with egg”.

RESULTS

A total of 21424 bird species records contain ing photographs from eBird (14668 records from 3277 checklists) and iNaturalist observations (6756 records) were examined for this study. The total breeding records corresponded only to 1.4% (n=302) of the total records examined, from which photographs or videos had sufficient evidence to assess their breeding status (Figs. 4-10; Online Appendix 2). From the total breed ing records, 220 records were unique to eBird, 69 records were unique to iNaturalist, and 14 records were shared between both portals. We were able to retrieve breeding records spanning 45 bird species, including waterfowl, shorebirds, waterbirds, and landbirds (Fig. 2), which consti tuted 37.8% of the 119 native and established introduced species that might potentially breed in the LMA (Municipalidad Metropolitana de Lima, 2020).

Fig. 4 Examples of records of copulation. For location details, see Online Appendix 2. (A) Zenaida meloda (8 March 2020, ML253545841, James Court); (B) Columbina cruziana (25 September 2017, ML107087521, Gord Smith); (C) Fulica ardesiaca (31 December 2020, iNat[Photos]108843442, Ruth Gutiérrez); (D) Larus dominicanus (1 November 2015, ML286328901, Oscar Johnson); (E) Parabuteo unicinctus (4 April 2020, ML220540281, Rutger Koperdraad); (F) Passer domesticus (13 September 2020, ML262861781, Ruth Gutiérrez). 

Fig. 5 Examples of records carrying nesting material. For location details, see Online Appendix 2. (A) Zenaida meloda (6 October 2019, iNat[Photos]54916985, Camden Bruner); (B) Fulica ardesiaca (27 July 2019, iNat[Photos]71812367, Mónica Paredes); (C) Troglodytes aedon (7 January 2018, ML81127221, Rutger Koperdraad); (D) Sicalis flaveola (3 April 2015, ML23543891, Laura Mae); (E) Volatinia jacarina (5 September 2015, iNat[Photos]2406656, Manuel Miranda); (F) Coereba flaveola (15 April 2011, ML79594031, Larry Silvio). 

Fig. 6 Examples of occupied nests. For location details, see Online Appendix 2. (A) Podiceps major (11 September 2016, ML70764301, Merryl Edelstein); (B) Zenaida auriculata (14 December 2019, ML193256191, Rutger Koperdraad); (C) Amazilis amazilia (4 February 2016, ML24650971, Laurie Koepke); (D) Fulica ardesiaca (21 November 2012, ML179817281, Simon Walkley); (E) Himantopus mexicanus (9 September 2019, ML235292521, Reid Rumelt); (F) Tachuris rubrigastra (11 September 2018, ML117175631, Craig Caldwell). 

Fig. 7 Examples of nests with eggs. For location details, see Online Appendix 2. (A) Zenaida meloda (24 April 2020, iNat[Photos]68437119, Katherine Zapata); (B) Charadrius vociferus (19 December 2020, ML289965911, Jorge Ubillas); (C) Haematopus palliatus (17 October 2020, ML272466331, Autoridad Municipal de Los Pantanos de Villa PROHVILLA). 

Fig. 8 Example of nestlings. For location details, see Online Appendix 2. (A) Haematopus palliatus (18 July 2019, ML169273161, Daniel Lane); (B) Burhinus superciliaris (14 February 2016, ML359146251, Santiago Pease); (C) Nycticorax nycticorax (29 June 2019, iNat[Photos]65401463, Tatiana Danilina); (D) Coragyps atratus (9 September 2020, iNat[Photos]141596082, Jose Huaman); (E) Parabuteo unicinctus (2 February 2019, ML349896001, Shirley Freyre); (F) Volatinia jacarina (31 December 2012, ML364482411, Cynthia Cerna). 

Fig. 9 Example of records feeding young. For location details, see Online Appendix 2. A) Podiceps major (19 September 2016, ML35580861, Andrew Spencer); (B) Amazilis amazilia (8 January 2017, ML351932971, Mariano Cordova); (C) Fulica ardesiaca (19 October 2019, ML182997981, David Belmonte); (D) Anthus peruvianus (4 October 2015, ML21069791, David Beadle); (E) Coereba flaveola (6 June 2016, ML36073761, Gil Ewing); (F) Dives warczewiczi (30 March 2013, iNat[Photos]141455481, Isabel Guerra). 

Fig. 10 Example of fledglings. For location details, see Online Appendix 2. (A) Oxyura jamaicensis (18 August 2019, ML174198501, Mónica Paredes); (B) Systellura decussata (13 July 2015, ML367546241, Biología Reproductiva de Aves Peruanas [Rodrigo Pulgar]); (C) Charadrius vociferus (13 December 2017, ML85015661, Joseph Morlan); (D) Camptostoma obsoletum (24 November 2010, ML96678041, Stephen Gast); (E) Tyrannus melancholicus (3 February 2019, ML140234891, Rutger Koperdraad); (F) Volatinia jacarina (1 November 2014, ML204817641, Oscar Delareina). 

Geographic distribution of breeding re cords and breeding activity of species. We recovered records from 27 districts of the LMA, including 24 districts from Lima Province and 3 districts from Callao Province (Fig. 1; Online Appendix 2). The amount of breeding data com ing from both citizen science portals together was markedly greater in the Chorrillos district (60.3%, n=182) than in other districts of the LMA. The American Oystercatcher (Haematopus palliatus) and the Great Grebe (Podiceps major) were among the top two species with most records contributed during our study period, represent ing 17.2% (n=52), and 11.9% (n=36) of the total breeding records found, respectively (Fig. 2). The West Peruvian Dove (Zenaida meloda) was the species with the highest diversity of breeding re cords compiled per species by showing six types of breeding activity records (Fig. 2).

Timing of breeding activity. We were able to establish breeding periods for 24.4% (n= 11) of the total breeding bird species retrieved from citizen science portals. The monthly frequen cy of breeding records for each species and pre liminary information about periods of breeding activity are shown in Table 1. Evidence of breed ing activity was found for the following species: the Ruddy Duck (Oxyura jamaicensis) and the Pied-billed Grebe (Podilymbus podiceps) during the humid-cold season; the Common Gallinule (Gallinula galeata) at the end of the humid-cold season and during the dry-warm season; the Peruvian Thick-Knee (Burhinus superciliaris), the Vermilion Flycatcher (Pyrocephalus rubinus), the Long-tailed Mockingbird (Mimus longicau datus), and the Saffron Finch (Sicalis flaveola) during the dry-warm season; and the American Oystercatcher (H. palliatus), the Great Grebe (P. major), the West Peruvian Dove (Z. meloda), and the Slate-colored Coot (Fulica ardesiaca) throughout the year.

Table 1 Monthly records of breeding activity per each species found in the LMA, retrieved from eBird and iNaturalist’s citizen science data from January 2000 to December 2020, compared to preliminary data for the Peruvian central coast. Dry-warm season occur from December to April, and Humid-cold season occur from June to October. Concerning preliminary data: ¹Coincident; ²Improved; ³Noncoincindent. 

After comparing our data with preliminary information, we determined that 31.1% (n=14) of the total number of breeding species in our data was “coincident”, and in the 13.3% (n=6) of these species our data was “noncoincident”.

In 55.6% (n=25) of these species, an “improved” knowledge of their timing of breeding was ob tained; including the fact that in 18 species, such as the White-cheeked Pintail (Anas ba hamensis), Ruddy Duck (Oxyura jamaicensis), the White-tufted Grebe (Rollandia rolland), the Pied-billed Grebe (Podilymbus podiceps), the West Peruvian Dove (Zenaida auriculata), the Common Gallinule (Gallinula galeata), Slate-colored Coot (Fulica ardesiaca), Kelp Gull (Larus dominicanus), Striated Heron (Butorides striata), Great Egret (Ardea alba), Black Vulture (Coragyps atratus), Harris’s Hawk (Parabuteo unicinctus), American Kestrel (Falco sparveri us), Tropical Kingbird (Tyrannus melancholi cus), Peruvian pipit (Anthus peruvianus), Scrub Blackbird (Dives warczewiczi), Yellow-headed Blackbird (Chrysomus icterocephalus), Saffron Finch (Sicalis flaveola), there was no preliminary data available about the timing of their breeding activity from central coast of Peru (Table 1).

Patterns of nesting behavior. Records of nesting birds represented 27.1% (n=82) of the total breeding records, from which we described following behavioral patterns:

Nesting-habitat use: We identified 23 species of nesting birds occurring in all three types of habitats (Fig. 3). Blue space had the greatest number of breeding species (n=10) and num ber of occurrence records (52.4%, n=43), from which nests were found in wetlands and sandy beaches. The American Oystercatcher (H. pal liatus) was the species with the highest number of records (n= 21) in this type of habitat. Green and gray spaces comprised nine and six breeding species and represented 24.4% (n=20) and 23.2% (n=19) of total number of occurrence records of nesting birds, respectively (Fig. 3). Nests found on green spaces were found on parks, and street trees, whereas nests on gray spaces were found on roof eaves, roof drainages, window box plant ers, indoor plant pots, shed bases, pillars, utility poles, electricity wires, and roadsides. The West Peruvian Dove (Z. meloda) was the species with highest number of records found in green (n = 5) and gray spaces (n=9).

Host plant preference: Records of host plants that we were able to identify represented 30% (n=18) of total records of nesting birds. We identified 15 species of plants that provided sup port to eight different species of nesting birds in two habitat types (Table 2). Nests were built on horizontal branches or forks of field-grown trees, and pot-grown plants located on green and gray spaces (Online Appendix 2).

Table 2 Nesting of different bird species observed on various plants by grown habit and habitat type found in the LMA, retrieved from eBird and iNaturalist’s citizen science data from January 2000 to December 2020, compared to preliminary data for the Peruvian central coast. Occurrence status of birds and plants: ¹Native species; ²Introduced species. Growth habit of plants: ^aField-grown tree found in green area; ^bPot-grown plant found in gray área 

Clutch size: Records of nest with eggs rep resented 26.8% (n=22) of the total nesting bird records and comprised three species occur ring in two habitat types (Table 3). Clutch size ranged from one to four eggs for the American Oystercatcher (H. palliatus) and three eggs for the Killdeer (Charadrius vociferus) in the blue space, and two eggs for the West Peruvian Dove (Z. meloda) in one record located in the green space.

Table 3 Clutch size of different bird species by habitat type found in the LMA, retrieved from eBird and iNaturalist’s citizen science data from Jan 2000 to Dec 2020, compared to preliminary data for the Peruvian central coast. 

DISCUSSION

We demonstrated that eBird and iNaturalist can be used together to study the breeding biol ogy of Peruvian birds. Our study is the first to determine breeding sites and patterns of breed ing activity of bird species in the LMA based on an in-depth review of citizen-science data, as well as, to confirm and improve the knowledge of timing of breeding activity and nesting patterns of birds from the Peruvian central coast. The study of the breeding biology of birds involving citizen participation originated in northwestern Europe (Atlas of breeding birds in Great Britain and Ireland; ^{Sharrock, 1976}) and this idea was subsequently spread to the rest of the world at a wide range of spatial scales (^{Gibbons et al., 2007}). In South America, the first effort was led by Chile, with its first national atlas (“Atlas de Aves Nidificantes de Chile”), which included the participation of nearly 1500 observers, who simultaneously collected and uploaded more than 600000 breeding records to eBird during 2011-2016 (Red de Observadores de Aves y Vida Silvestre de Chile, 2018). However, obtaining nationwide or even city-wide data for monitor ing breeding activity of birds can be very time-consuming and costly. Hence, the thorough ex amination of already existing citizen science data constitutes a less time-consuming and expensive method to obtain valuable information on breed ing birds. Also, our data was collected without concerns of causing an additional disturbance to the nest sites or damaging the health of captured birds. Further strengths of this study are that the information gathered from photographs can be preserved for studies on repeatability or fu ture research uses. However, conclusions should be taken cautiously owing to spatial, taxonomic, and temporal limitations in citizen science data that will be further discussed below.

Breeding records were concentrated to wards the southwest of the LMA, specifically in the Chorrillos district, in the Pantanos de Villa wetlands (Fig. 1, Online Appendix 1), the latter being an area that provides refuge for 211 bird species, and breeding ground for 61 resident bird species (^{Pulido, 2018}). However, the considerably greater number of breeding records might not only be attributed to the occurrence of birds in this important ecological reserve but also to their high and increasingly growing demand for bird-watching tourism in recent years (^{Carhuas & Jacinto, 2020}; ^{Aponte et al., 2020}). On the other hand, it is often assumed that the occurrence of a species at a given location during the breeding season is an indication of breeding activity and suitable breeding habitat. However, it might not always be the case. Survey sites might include non-breeding individuals, juveniles away from natal sites, might not entirely cover the nesting area, or the detection of unpaired males by sight might be more related to the presence of relegat ed males looking for sites in low-quality habi tats for breeding (^{Fogarty et al., 2022}; ^{Gorleri & Areta, 2021}). The location of records shown here does not necessarily reflect breeding habitat suit ability. For example, although most of the breed ing records for the American Oystercatcher (H. palliatus) were gathered from the Pantanos de Villa wetlands, previous research has evidenced low reproductive success for this species during the summer-fall season caused by anthropogenic disturbances in such areas (^{Arenas et al., 2020}). In conjunction with future contributions, how ever, these data should help establish suitable habitats for breeding birds in LMA. Nowadays, this type of study can benefit from associations with citizen science crowdsourcing campaigns to cover spatial limitations in photograph/video distributions.

Despite such diversity of breeding birds, our records can be taxonomically biased because of an uneven detection of breeding bird taxa. Often breeding birds are secretive when they are on the nest to avoid attracting attention from pre dators, and some species are far more secreti ve than others; hence, breeding individuals of some species might not be easily approached or detected by an observer and, consequently, are frequently overlooked. Such is the case of the Peruvian Pipit (Anthus peruvianus) and the Tschudi’s Nightjar (Systellura decussata), two secretive birds recently reclassified to species level and uncommonly to commonly distributed along the Pacific coast of Peru (^{Begazo, 2020}); being the latter zone an area increasingly recog nized as an avian region of endemism and from which very little is known about the biology of its species (^{Arcco et al., 2020}). A. peruvianus oc curs in open areas often hiding in grassy riparian zones, while S. decussata is a nocturnal species that occurs in arid scrubs or urban sites (del Hoyo et al., 2020a, b). Our compilation of citizen science data has revealed the first and second photographic evidence of breeding activity of S. decussata and A. peruvianus for Peru (Figs. 9F, 10B, Online Appendix 2), respectively; although we do not discard the existence of other photo graphic evidence for these species that remain unpublished. It is also important to note that the record of S. decussata consisting of the presen ce of a downy fledgling accompanied by an adult inside an uninhabited area of a house building (Fig. 10B) evidences how observations provided by building owners can be of great benefit becau se they may encounter species that might be very difficult to detect elsewhere.

Our results also provided information about timing of breeding activity for 45 bird species, from which we were able to establish breeding pe riods for 24.4% (n=11) of these species. However, because of the relatively small sample size per species and the arbitrary nature of dates when observations were made, records presented here do not strictly reflect the presence or absence of breeding seasonality. Moreover, despite the coin cident and improved knowledge of breeding activ ity timing gathered for 86.7% (n=39) of total spe cies described here, it was not possible to estab lish periods of breeding activity for 53.3% (n=24) of them due to the lack of continuous months of breeding activity. However, preliminary informa tion did allow us to establish a clearer period of breeding activity for the Vermilion Flycatcher (Pyrocephalus rubinus) and other 10 additional species (Table 1): the Cinnamon Duck (Spatula cyanoptera) during the humid-cold season; the Wren-like Rushbird (Phleocryptes melanops) and the Many-colored Rush Tyrant (Tachuris rubrigastra) during the humid-cold season and at the beginning of the dry-warm season; the Rufous-collared Sparrow (Zonotrichia capensis) during the dry-warm season with an additional short period during the humid-cold season; the Killdeer (Charadrius vociferus) during the dry-warm season; and the Croaking Ground-Dove (Columbina cruziana), the Vermilion Flycatcher (Pyrocephalus rubinus), the House wren (Troglodytes aedon), the House Sparrow (Passer domesticus), the Blue-black Grassquit (Volatinia jacarina), and the Bananaquit (Coereba flaveola) throughout the year. The latter fact brings the number of species able to breed year-round to 10, representing 47.6% of species with known breed ing periods in our study. This, coupled with the existence of more discrete periods of reproductive activity in most of these species in further north or south latitudes of their distributions, includ ing the American Oystercatcher (H. palliatus), Great grebe (P. major), Vermilion flycatcher (P. rubinus), House Wren (T. aedon), House Sparrow (P. domesticus), Bananaquit (C. flaveola), and the Blue-black Grassquit (V. jacarina) (^{Greenquist, 1982}; ^{Wunderle, 1982}; ^{Bugoni et al., 2002}; ^{Zuria & Rendón-Hernández, 2010}; ^{Ippi et al., 2012}; ^{Marini et al., 2012}; ^{Llambías et al., 2015}; ^{Dubois, 2016}; ^{Figueroa & Stucchi, 2016}; Medrano et al., 2019; ^{Hilty & Christie, 2020}; ^{Johnson, 2020}; ^{Llimona et al., 2020}; ^{Lowther & Cink, 2020}; ^{Rising, 2020}), provides additional evidence that tropical birds can breed throughout the year (^{Hau et al., 2008}; ^{Echeverry-Galvis & Córdoba-Córdoba, 2008}). Further continuous records over multiple years are needed to better understand the temporal patterns of breeding activity for these species.

Concerning nesting habitat selection, we found species nesting in all three types of habi tats, with the lowest number of species and oc currences found in gray spaces. However, be cause of the small sample size of nesting birds in each type of habitat, it is difficult to envision how habitat selection influences the choice of nesting territory in the LMA. It is known that introduced bird species tend to exploit anthropo genic habitats that are inefficiently used by na tive species such as the gray spaces (^{Savard & Falls, 1981}; ^{Sol et al., 2012}); however, there are other important factors that can influence nest ing habitat selection of birds in urbanized areas, including vegetation coverage, nest protection, and the level of human disturbance (^{Pennington & Blair, 2011}; ^{Soulsbury & White, 2015}; ^{Zhou et al., 2020}). The presence of several native species nesting on gray spaces, including, among others, the Peruvian Thick-knee (B. superciliaris) and the West Peruvian Dove (Z. meloda), might sug gest a limited availability of suitable nesting sites given the growing reduction of green and blue spaces in Lima city (^{Quispe, 2017}; ^{Velásquez et al., 2018}). Gray spaces can also offer protection against bad weather, brood parasitism, and nest predation as has been shown for urban birds in other parts of the world (^{Liang et al., 2013}; ^{Vincze et al., 2017}; ^{Mainwaring, 2015}). However, the main disadvantage of gray spaces as nesting habitats is that they sometimes can act as eco logical traps because species can be negatively affected by the temporary availability of man made structures which may cause an inadver tent loss of nesting sites (^{Reynolds et al., 2019}). Concerning host-plant preference of nesting birds, most of the plant species identified were considered as introduced while only one were na tive to the LMA (Table 2). Introduced species in cluded nine field-grown tree species and six pot-grown plants, which provided nesting support to five native and two introduced bird species in the LMA. The only native host plant species identified was a field-grown tree that provided nesting support to one native bird species in the LMA. Previous studies have evidenced that some native birds prefer to nest under the protective cover of introduced plant species, however, in some cases, introduced plants can also act as ecological traps by altering vegetation structure, decreasing food availability, and increasing nest predation (^{Schmidt & Whelan, 1999}; ^{Borgmann & Rodewald, 2004}; ^{Nelson et al., 2017}). More evi dence on plant-nesting bird interaction is needed to understand the impact of such introduced plant species on the bird breeding community at our study site. Furthermore, host plant spe cies only coincided for two species with previous reports from the Peruvian central coast, which is likely influenced by the lack of information on this aspect of avian breeding biology in such region and/or the cultivated nature of many plant species occurring in the LMA. Vegetation structure in urban cities depends on residents’ aesthetic values, protection, and economics, as well as jurisdictional policies, and the LMA is no stranger to such conditions (^{Zhang et al., 2013}; ^{Sabogal, 2021}; ^{Handayani & Mardikaningsih, 2022}). Concerning clutch size, the three native bird species from which we were able to appreci ate nest content, namely the Killdeer (C. vocifer us), the American Oystercatcher (H. palliatus), and the West Peruvian Dove (Z. meloda), showed invariable clutch sizes to those reported for these species in other areas of the Peruvian central coast. Unfortunately, we did not have enough data to infer the potential influence of urbaniza tion on clutch size or other aspects of nesting be havior. For example, it has been evidenced that increased urbanization correlates significantly with smaller clutch size, lower offspring productivity, and increased nestling periods (^{Ibañez-Álamo & Soler, 2010}).

On the other hand, citizen science portals might not constitute a good approach to study some other aspects of the breeding biology of birds that still need to be addressed through in situ field observations or other methods. Detailed descriptions of nest building, nestling growth, nest survival rates, regimes of parental care, or nest-site fidelity cannot be inferred from a remote examination of photographs made by a third-party investigator. Moreover, assessing age, sex, and reproductive status of individuals from the photographs described here might not be comparable in accuracy to examining plum age, sexing on the bases of external character istics (i.e., brood patch or cloacal protuberance), gonad descriptions, or other extended pheno types (e.g., eggs, nest, etc.) retrieved from living birds or museum specimens. The combination of all these methods, however, will allow us to fill multiple gaps related to the breeding biology of birds with more accuracy and from a broader perspective. Furthermore, citizen-science data can also provide complementary information to answer questions related to other aspects of the natural history of birds, including distributional patterns, foraging ecology, molt patterns, and habitat use (^{McCaffrey, 2005}; ^{DeGroote et al., 2020}; ^{Panter & Amar, 2021}; ^{Pyle, 2022}; ^{Lopes & Schunck, 2022}). Hence, we encourage research ers to gather and publish similar or larger-scale studies to those presented here for other urban and non-urban areas of Peru and other neotropi cal developing countries. Finally, we encourage the participation of citizens in contributing with more photographs of birds to citizen-science por tals to increase scientific knowledge and commu nity awareness of the value of bird biodiversity and their habitats.

Supplementary information online. http://revista.macn.gob.ar/ojs/index.php/RevMus/rt/suppFiles/769/0