RIS313049 Bijlage 2

Received: 15 September 2021

Accepted: 21 December 2021

DOI: 10.1002/2688-8319.12128

RESEARCH ARTICLE

Nutrient fertilization by dogs in peri-urban ecosystems Pieter De Frenne1 Pieter Vangansbeke1

Mathias Cougnon2,3

Geert P. J. Janssens4

1

Forest and Nature Lab, Department of Environment, Faculty of Bioscience Engineering, Ghent University, Gontrode, Belgium

2

Plant Science Unit, Research Institute for Agriculture, Fisheries and Food (ILVO), Melle, Belgium

3

Department of Plants and Crops, Faculty of Bioscience Engineering, Ghent University, Melle, Belgium

4

Department of Veterinary and Biosciences, Faculty of Veterinary Medicine, Ghent University, Merelbeke, Belgium

Abstract 1. (Semi-)natural ecosystems provide many important benefits to nature and people, but are often located near populated and urbanized areas across the globe. During recreational activities, many people bring dogs into peri-urban forests and nature, but their nutrient inputs per unit space and time via dog faeces and urine into ecosystems remain scarcely quantified. 2. Here, we estimate net fertilization rates of dogs in peri-urban ecosystems, with a focus on nitrogen (N) and phosphorus (P) because of their evident effects on plant biodiversity. We used 487 direct-count censuses over 1.5 years to collect accurate dog abundance data per hectare per year in four sites in peri-urban forests and

Correspondence Pieter De Frenne, Forest & Nature Lab, Department of Environment, Faculty of Bioscience Engineering, Ghent University, Geraardsbergsesteenweg 267, 9090 Gontrode, Belgium. Email: Pieter.DeFrenne@UGent.be Handling Editor: Shinichi Tatsumi

nature reserves in Belgium. Based on estimated dog densities and a systematic literature search of nutrient concentrations in urine and faeces, we calculate N and P fertilization rates from urine and faeces deposits, also propagating uncertainty and variability in these estimates. 3. We find that canine N and P fertilization rates on average amount to 11 kg N (more or less equally from urine and faeces) and 5 kg P (predominantly from faeces) per hectare per year, respectively. These estimated amounts are substantial when compared to atmospheric inputs of N and extractable amounts via traditional nature management (e.g. mowing and hay removal). 4. Our estimated dog N and P fertilization rates in peri-urban forests and nature are substantial. Such levels of nutrient inputs may considerably influence biodiversity and ecosystem functioning, and co-determine restoration outcomes.

Our results underpin the need for managers and policy makers to more often (i) consider currently neglected nutrient inputs by dogs in management plans and restoration goals, (ii) communicate to dog walkers the role of their dog as ‘fertilizer’ and highlight the necessity to remove at least canine solid faecal waste, (iii) in sensitive oligotrophic ecosystems with species adapted to nutrient-poor soils, establish nearby off-leash dog parks, enforce the use of short leashes and/or apply dog bans such that high dog abundances can be avoided.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2022 The Authors. Ecological Solutions and Evidence published by John Wiley & Sons Ltd on behalf of British Ecological Society Ecol Solut Evid. 2022;3:e12128. https://doi.org/10.1002/2688-8319.12128

wileyonlinelibrary.com/journal/eso3

1 of 9

2 of 9

DE FRENNE ET AL .

KEYWORDS

biodiversity loss, Canis familiaris, eutrophication, fertilization, forest and nature management, nitrogen deposition, nutrient enrichment, peri-urban ecosystems, phosphorus, recreation

1

INTRODUCTION

N and P inputs have been shown to strongly negatively impact biodiversity and ecosystem function (Bobbink et al., 2010). In plant com-

Peri-urban ecosystems such as forests, (semi-)natural grasslands, wet-

munities, for instance, N addition decreases species richness in a wide

lands and heathlands in populated areas across the globe provide many

range of ecosystems (De Schrijver et al., 2011), whereas P fertilization

important benefits to nature and people such as biodiversity conser-

eradicates the niche of many threatened species (Wassen et al., 2021).

vation, carbon drawdown, nutrient cycling, climate regulation, wood

In many ecosystems, also in populated areas, forest and nature man-

and food production and recreation (Perring et al., 2013). In many

agement is specifically directed towards lowering soil nutrient concen-

instances, such ecosystems are on the one hand of important con-

trations via practices such as mowing with hay removal, local topsoil

servation concern, but on the other hand also experience significant

removal and phytoextraction (sometimes also referred to as mining)

amounts of daily human visitors, especially when located near rela-

(Pegtel et al., 1996; Schelfhout et al., 2015, 2017, 2019). Neglecting

tively densely populated and urbanized areas. The impacts of human

the nutrient inputs from dogs in such cases might result in an under-

recreationists on disturbance of wildlife such as breeding birds are rel-

estimation of the time needed for ecological restoration and the costs

atively well-quantified (Arnesen, 1999; Lenth et al., 2008). However,

involved. Misinformed restoration advice might negatively affect bio-

many people also bring domestic dogs (Canis familiaris) on recreational

diversity and the associated ecosystem services.

activities. While effects of dogs on wildlife via direct mortality, distur-

Here, we quantified N and P inputs from canine urine and faeces in

bance and disease transmission have been relatively well documented,

peri-urban forests and nature reserves specifically managed for biodi-

their fertilization effects have received considerably less attention

versity conservation and consisting of small forest patches, wetlands

(Weston et al., 2014).

and grasslands with vulnerable, species-rich vegetation. Innovative to

There are an estimated 87 million dogs in Europe and 72 million in

our approach is that we used nearly 500 dog density transect counts

the United States (FEDIAF, 2019; Paradeis et al., 2013). In Europe and

across a time span of 1.5 years to estimate N and P inputs. Combined

the United States, about 25% and 49% of households owns at least one

with a systematic review of dog urine and faeces N and P concentra-

dog, respectively (Allen et al., 2020; FEDIAF, 2019). Via their urine and

tions, this approach enabled us to calculate dog densities and fertiliza-

solid waste (faeces), dogs bring in significant amounts of nutrients into

tion rates from both urine and faeces per unit space and time across the

ecosystems but this disturbance and its associated effects on biodiver-

peri-urban ecosystems.

sity have been often neglected so far. Dog faeces and urine count as net inputs, because dogs are fed at home with a protein-rich diet, in contrast to grazing cattle (Bos taurus L.), sheep (Ovis aries L.) or foraging

2

METHODS

birds that feed off the land and recycle nutrients within the ecosystem. While several studies have detected significantly elevated soil nutri-

2.1

Study area

ent concentrations in areas with many dogs (Allen et al., 2020; Bonner & Agnew, 1983; Oates et al., 2017; Paradeis et al., 2013), dog fer-

This study was conducted near Ghent, a medium-sized city (about

tilization rates per unit time and space (kg per ha per year), however,

260,000 inhabitants) in Belgium with a temperate climate (mean

have not been quantified at the ecosystem level such that manage-

annual temperature of 10.3◦ C and mean annual precipitation of

ment actions with regard to dogs tend to only focus on their effects on

789 mm between 1970 and 2000; Fick & Hijmans, 2017). Atmospheric

wildlife.

N deposition was 22.7 kg N ha−1 year−1 in 2019 in the study area

Nutrient inputs from canine urine and faeces can have important

(Flemish Environmental Agency, 2020). We selected four study sites

effects on soil nutrient concentrations, particularly in terms of the

in peri-urban nature reserves less than 5 km from the city centre (Fig-

macronutrients nitrogen (N) and phosphorus (P). In areas with a lot

ure 1). The study sites are popular for recreation but also hold impor-

of dog walkers, and especially near walking paths, elevated soil P and

tant biodiversity values. The study sites differ in size, in vegetation

N concentrations are found and stable isotope analyses confirmed

type, in management and in accessibility (Table 1), with visitors in study

dogs as the source (Allen et al., 2020; Bonner & Agnew, 1983). These

sites 1, 2 and 3 restricted to trails but without physical boundary to the

patterns were still apparent even 3 years after an imposed dog ban

vegetation and with visitors in site 4 legally permitted to leave the trails

(Bonner & Agnew, 1983). Paradeis et al. (2013) also detected strong soil

and walk freely in the reserve. All study sites are part of larger nature

nutrient and pH impacts of dog urine within off-leash dog parks. Finally,

reserves and were delineated based on the physical ability of dogs to

also on marine recreational beaches, dog faeces can result in signifi-

enter when off leash (borders of the study sites were often demarcated

cant nutrient inputs and marine pollution (Oates et al., 2017). Elevated

by rivers, fences or roads).

3 of 9

DE FRENNE ET AL .

F I G U R E 1 Study area. (a) Location of Ghent, Belgium in Europe (red circle). (b) Location of the four study sites around the city centre of Ghent. Numbers of sites (1-4) also refer to Table 1. (c) Detailed map of the four study sites (black dashed lines) and the covered direct count transects (red line) TA B L E 1 Overview of the characteristics of the four study sites: Location (numbers refer to the map in Figure 1), area, dominant vegetation type, management and access

Study site

Location (latitude, longitude)

Area (ha)

1. Vinderhoutse Bossen

51.08◦ N, 3.65◦ E

18.4

Alluvial forest (H91E0)

No intervention management

Paths

2. Meerskant, Bourgoyen

51.07◦ N, 3.66◦ E

5.9

Calthion grassland (H6410)

Mowing with hay removal

Paths

3. Spoorwegberm, Bourgoyen

51.06◦ N, 3.68◦ E

3.3

Lowland hay meadow (H6510)

Mowing with hay removal

Paths

4. Gentbrugse Meersen

51.04◦ N, 3.79◦ E

49.4

Oak-hornbeam-forest (H9160) and lowland hay meadow (H6510)

Low-density livestock grazing

Entirely accessible, no need to stay on paths

2.2

Dominant vegetation type (with Natura 2000 code)

Dog counts

Current management

Access restricted to

2017). Transect counts were executed two to four times weekly in each site, regardless of weather and at varying times throughout the day. In

Dogs were counted along transects in the four study sites between

total, 487 counting events took place, more or less spread throughout

February 2020 and June 2021 for a total of 487 censuses. We fol-

the week: 46 counts on Mondays, 66 on Tuesdays, 76 on Wednesdays,

lowed a modified direct-count census to accurately and representa-

82 on Thursdays, 71 on Fridays, 61 on Saturdays and 85 on Sundays. No

tively quantify dog presence in each reserve (Oates et al., 2017). The

permission was needed for this fieldwork.

transects were laid out in a way that the whole study site could be inspected when the transect was covered. A single observer per study area recorded all unique dogs on and off-leash (recorded separately)

2.3

Nutrients in urine and faeces

while covering the transect at a constant speed. By accounting for the size of the study site and by assuming a mean presence of the dogs of

For the nutrient concentrations of canine urine and faeces, we per-

1 h in the larger study sites 1 and 4, and of half an hour in the smaller

formed a systematic literature search and used the mean and vari-

study sites 2 and 3 and a mean daylength of 12 h, the data of every cen-

ation across the primary studies (Table S1). We searched for stud-

sus were expressed as a number of dogs per ha per day (cf. Oates et al.,

ies in Web of Science using the keywords ‘dogs and (phosph* or

4 of 9

DE FRENNE ET AL .

nitrogen) and (digestib* or excretion) and (urinary or f?ecal)’ in early

and the repeated measurements within each site (time series) was

Nov. 2021. This search resulted in 180 potentially suitable studies.

hence taken into account.

Those 180 papers were then manually screened for studies that met

Second, since there are several factors that can lead to overestima-

the following criteria: (i) N and/or P concentrations of dog urine and/or

tion or underestimation of our inferred fertilization rates (e.g. variation

faeces were reported or could be calculated from available data; (ii)

in nutrient concentrations as a result of dog food quality and quantity,

if treatments of diets or diseases were reported, we only included

imperfect detection of dogs during transect census counts, the amount

the control treatments and diets that could be considered as com-

of urine or faeces deposited as dependent on dog size, walk duration,

mon practice. Nutrient concentrations were obtained either (i) directly

dog size distribution and faeces collection rates), we also propagated

if the concentrations were mentioned in the original papers or (ii) if

uncertainty and variability on parameters as a second step. There-

digestibility of N (or crude protein) or P was reported, faecal concentra-

fore, we resampled 999 bootstrap samples from the 487 censuses in

tions were calculated based on food intake, dry matter concentration

the different study sites and implemented the estimated mean and

in the diet, dry matter concentration of the faeces, dietary N or P con-

standard deviation from the nutrient concentrations obtained in the

centrations and digestibility coefficients, according to the principle of

literature review (Table S1). For the parameter values for which no lit-

apparent digestibility calculations: Digestibility of N (or P) (%) = 100 −

erature estimates were available (urine volume, faeces mass and dog

100 × [(faeces (g) × faeces N (or P) concentration (g/kg)) / (food (g) ×

residence time), we calculated the standard deviation as a value of 20%

food N (or P) concentration (g/kg))]. Finally, faecal N concentrations

of the mean to obtain a normal distribution of estimates. For these

were based on 19 diets from six studies (Beynen et al., 2002; Cargo-

bootstrapped estimates of the total N and P inputs per ha per year, we

Froom et al., 2019; De Smet et al., 1999; Forster et al., 2012; Pinna et al.,

then report the mean and 5 and 95 percentiles of the distribution. The

2018; Wood et al., 2004), whereas faecal P concentrations were based

variability in the model parameters is shown in Figure S1. All data anal-

on 17 diets from five studies (the same, except Forster et al., 2012). Uri-

yses were executed in R version 4.0.4 (R Core Team, 2021) and graphs

nary N concentrations were based on two studies (Beynen et al., 2002;

produced with the ggplot2-library (Wickham, 2016).

Castrillo et al., 2001), whereas urinary P concentrations were based on 10 diets from three studies (Atwal et al., 2021; Stevenson et al., 2003; Wood et al., 2004), including one very well-documented study (Atwal

2.5

Scenario analysis

et al., 2021). Because dogs are carnivores and mainly fed with a protein-rich diet,

To investigate the effect of dog owner behaviour on N and P deposition,

the nutrient concentrations in urine are relatively high. Urine N and P

we also modelled the effect of owners keeping all dogs on a short leash

concentrations amount on average to 18.7 g N L−1 and to 484.6 mg

(2 m) and collecting all solid faecal waste (not possible with urine), as

P L−1 (Table S1). Cattle urine, for comparison, has typical N concentra-

actually prescribed by the current legislations in the different nature

tions of 0.7–10.2 g N L−1 (Hoogendoorn et al., 2010). We adopt canine

reserves (note there is no legal limit to leash length). If all dogs are kept

faecal concentrations of 44.3 mg N g−1 faecal dry mass and 32.0 mg

on a leash of 2 m, the area of the fertilized zone is strongly reduced for

P g−1 faecal dry mass (Table S1). These concentrations are, again for

study site 1 (reduced to 0.744 ha), study site 2 (0.317 ha) and study site

comparison, higher than N and P concentrations of 10–30 mg N g−1

3 (0.348 ha). In this scenario analysis, we did not consider study site 4,

and 1–4 mg P g−1 reported for savanna ruminants (Sitters et al., 2014)

because visitors there are legally permitted to leave the trails and are

and P concentrations of cattle, deer and sheep dung which ranges

allowed to roam freely with dogs on a leash; the disturbed area thus

between 5.5 and 8 mg P g−1 (McDowell & Stewart, 2005). For solid

remained the same.

waste (faeces), we assume that each dog produces faeces once on each trip with a mean dry scat weight of 100 g (de Molenaar & Jonkers, 1993). For urine, we assume that each dog deposits one quarter of the

3

RESULTS

daily 736 ml urine volume production per day (Beaver, 1999; Paradeis et al., 2013) during a walk to a dog park, and thus that 184 ml urine is

3.1

Dog densities

deposited per dog walk in the nature reserves. Across the four study sites and 487 count events, we counted 1629 dogs. We calculate a mean dog density of 1.3 dogs ha−1 day−1 off leash

2.4

Data analyses

and a mean of 2.9 dogs ha−1 day−1 on leash for a total estimated dog density of 4.2 dogs ha−1 day−1 (Figure 2). This is the equivalent of 1530

2.4.1

Nutrient deposition modelling

dogs ha−1 year−1 . There was significant among-site variation in dog densities as well as in leash use. Dog densities were highest at site 3

We estimated the annual deposition of N and P through urine and fae-

(which has a nature target value as species-rich grassland) where we

ces with an intercept-only mixed-effect model using the lme-function

counted a mean dog density of not less than 11.0 dogs ha−1 day−1 .

from the nlme-library (Pinheiro et al., 2021) with the day (numeric,

Overall, 66% of encountered dogs was on leash and 34% off leash.

counting the days from the first measurement) and site (four levels)

[Correction added on 7 February 2022 after first online publication:

as random-effect terms and including a temporal autocorrelation term,

percentages have been updated from 68% and 32% to 66% and 34%.]

with a continuous time covariate. The hierarchical nature of our data

Yet, the proportion of off-leash dogs strongly varied among reserves

5 of 9

DE FRENNE ET AL .

12.5

Number of dogs per hectare per day

10.0

site

7.5

1 2 3 5.0 4

2.5

0.0 off

on

sum

Off vs on leash

F I G U R E 2 Estimated dog densities (number ha−1 day−1 ), off and on leash, and their summed contributions, across the four study sites. Error bars denote standard errors. The grey dashed lines represent the mean across the four study sites

(most probably as a result of different management, policies and

removed using, for example, disposal bags, urine-only inputs amount to

enforcement), from merely 8% in site 1 to 29%–49% in the other sites

76.6 ± 27.8 kg N ha−1 year−1 and 2.0 ± 0.7 kg P ha−1 year−1 , that is a

(Figure 2). [Correction added on 7 February 2022 after first online pub-

reduction of 56% of N deposition and 97% of P deposition.

lication: percentages have been updated from 9% and 27%–52% to 8% and 29%–49%.]

3.2

Estimated annual fertilization rates

Based on the estimated dog densities and intercept-only mixed-effect modelling considering temporal autocorrelation, we estimate overall N and P inputs from faeces to amount to 6.5 ± 3.7 kg N ha−1 year−1 and 4.7 ± 2.7 kg P ha−1 year−1 , respectively. Urine-based inputs of N and P amounted to 5.0 ± 2.9 kg N ha−1 year−1 and 0.13 ± 0.07 kg P

3.4 Uncertainty and variability in model parameters: Bootstrap approach The resampling approach across 999 bootstraps samples propagating variation and uncertainty into our estimates of N and P concentrations of urine and faces, urine volume, faeces mass and dog residence times resulted in mean inputs of 12.2 kg N ha−1 year−1 (with 5 and 95 percentiles of 0.0 and 63.2 kg N) and 5.2 kg P ha−1 year−1 (with 5 and 95 percentiles of 0.0 and 24.1 kg P) (Figure S1).

ha−1 year−1 . The estimated total input of N and P across the four study sites is then 11.5 ± 6.5 kg N ha−1 year−1 and 4.8 ± 2.7 kg P ha−1 year−1 .

4

DISCUSSION

There was again significant among-site variation driven by the variation in estimated dog densities with maximum inputs of 31.3 kg N ha−1 year−1 and 13.1 kg P ha−1 year−1 at site 3 (Figure 3).

4.1 Fertilization by dogs is substantial and non-negligible Dogs appear to be a non-negligible, substantial and underestimated

3.3 Scenario analysis: What if all dogs are on leash and faeces is removed

source of nutrients into peri-urban ecosystems. Dog N input was

Finally, we analysed a scenario in which all detected dogs are on leashes

P input was 4.8 kg P ha−1 year−1 across all sites, with a peak of 13.1 kg

of maximum 2 m length (excluding study site 4, cf. Section 2). Nutrients

P ha−1 year−1 in the site with most dogs. Our estimates become

are then deposited in a significantly smaller area and concentrated in

even more significant when compared to (i) the potential annual

the near vicinity of the trails. This then leads to N and P deposition val-

nutrient removal rates with mowing and hay removal (traditional

ues of 175.3 ± 63.5 kg N ha−1 year−1 and 73.2 ± 26.5 kg P ha−1 year−1

management in semi-natural grasslands) that amount to 10–70 kg

within a zone 2 m left and 2 m right of each path (values again esti-

N and 2–20 kg P ha−1 year−1 in grasslands (Oelmann et al., 2009;

mated from intercept-only mixed-effect models). If the faeces would be

Schelfhout et al., 2015) and (ii) atmospheric N deposition inputs

11.5 kg N ha−1 year−1 across all sites, with a peak of 31.3 kg N ha−1 year−1 in the study site with the highest dog densities. The dog

6 of 9

DE FRENNE ET AL .

Phosphorus

Nitrogen

Deposition (kilogram per hectare per year)

15

30

10

site 20

1 2 3 4

5 10

0

0 faeces

urine

sum

faeces

Source of nutrients

urine

sum

Source of nutrients

F I G U R E 3 Estimated inputs (kg ha−1 year−1 ) of phosphorus (P) and nitrogen (N) via dog fertilization as faeces and urine, and their summed contributions, across the four study sites. Error bars denote standard errors. The grey dashed lines represent the mean across the four study sites

(5–25 kg N ha−1 year−1 across most of Europe; based on EMEP data in

volatilization losses and will become more gradually available for plants

Staude et al., 2020).

compared to urine. These nutrients will only affect the vegetation in the

Unlike atmospheric N or fertilizer N and P in grasslands under agri-

direct neighbourhood of the place where the faeces was deposited: for

cultural use, nutrients deposited by animals are not deposited uni-

instance, cattle dung pats covering 0.05 m2 affected grass growth in an

formly, but in patches. Carnivores, for instance, have the tendency to

area of about 0.25 m2 surrounding the dung and can have a measurable

deposit faeces on (aboveground) landmarks, for example near entrance

effect on grass growth for up to 2 years (Lantinga et al., 1987).

gates and trail intersections (Oates et al., 2017). In the case of dogs, urine deposition is patchy, even when only 1 m away from park paths (Allen et al., 2020). Half (44%) of the N deposited by dogs is via their

4.2

Effects on biodiversity

urine. Within urine patches deposited by grazing sheep and cattle, the N loads are in the range of 500–2000 kg N ha−1 . The N deposit in a

It is clear that the levels of fertilization by dogs estimated here can

dog’s urine patch is expected to be even higher as dog urine (18.7 g

potentially exert negative effects on biodiversity and ecosystem func-

N L−1 ) (Table S1) is more concentrated than cow (0.7–10.2 g N L−1 )

tioning of species-rich vegetation that are often pursued in forest and

or sheep urine (1.4–6.1 g N L−1 ) (Hoogendoorn et al., 2010). The N in

nature management. Higher nutrient levels lead to increased plant

urine has a very rapid effect on vegetation: within 2 days after excre-

growth, mostly by a limited number of nutrient-demanding species that

tion, it is completely transformed in plant available forms of N (ammo-

will outcompete specialists, particularly by taking away the available

nia and nitrate) (Lantinga et al., 1987). The N deposited in urine patches,

light (Hautier et al., 2009), causing plant species loss (De Schrijver et al.,

however, is prone to losses through volatilization (NH3 ) or leaching

2011) and homogenization of plant communities (Staude et al., 2020).

(NO3 − ) depending on vegetation, soil type, temperature and precipi-

This well-known effect of N pollution on vulnerable ecosystems has

tation. Research on the fate of urine N of grazing dairy cows shows that

led to the concept of critical deposition loads, which is defined as the

the proportion of the N recovered in the herbage varies between 58%

limit (‘effect threshold’) above which habitat quality risks to be signif-

and 32% for spring and autumn applied urine, respectively (Decau et al.,

icantly damaged by the impact of N deposition (Bobbink et al., 2010;

2003). An important but unknown part of the N deposited in winter and

Wamelink et al., 2021). For the vegetation types of our study sites 1–3,

autumn will not be taken up by the vegetation in the urine patches but

this critical deposition load is 20 kg N ha−1 year−1 , whereas it ranges

is prone to leaching and volatilization.

for study site 4 between 20 and 34 kg N ha−1 year−1 depending on

Only 3% of the total P deposited by dogs is via their urine. Unlike

the vegetation type (Van Dobben et al., 2012). With a current atmo-

N, P in the soil is much less mobile and will become gradually avail-

spheric N deposition of 5–25 kg N ha−1 year−1 across most of Europe

able to plants in the next growing seasons (Jarvis, 2000). The P and

(Staude et al., 2020), it is clear that the estimated canine N input of

N deposited through faeces thus represent 97% and 56% of the total

11.5 kg N ha−1 year−1 can have an important additional impact. Specif-

deposited P and N, respectively. This portion is less prone to leaching or

ically, within the urine patches N deposition has a strong effect on

7 of 9

DE FRENNE ET AL .

plant biodiversity and ecosystem processes (e.g. carbon and nutrient

enrichment (see Oates et al., 2017 for a discussion of other, more

cycling) on a microscale. On-site management such as mowing with hay

expensive management options). In addition, removing dog faeces pre-

removal can compensate much of the negative effects of N deposition,

vents the infection of grazing animals with zoonotic diseases, such as

but mostly fails to reduce the ecosystem N levels due to the constant

Neospora caninum. Dogs are the definitive hosts of this obligate intra-

input through deposition and is relatively expensive (Jones et al., 2016).

cellular parasite, but many other animal species can get infected. In wild

It is highly questionable that on-site management can also compensate

ruminants like roe deer (Capreolus capreolus L.) but especially domesti-

the negative effects of N deposition in dog urine patches, given the

cated grazers like cattle and sheep, infection with Neospora is a main

much higher concentrations compared to more uniform atmospheric

cause of abortion (Almería, 2013). Our findings also underpin that a

deposition.

‘stick and flick’ strategy to reduce the nuisance of treading in dog faeces

Also, excess P, most often due to former agricultural fertilization,

(as currently considered by, e.g., the Forestry Commission in Britain)

has a well-known negative effect on plant species richness (Ceulemans

is to be avoided. Second, keeping the dogs leashed (short leashes of

et al., 2014; Schelfhout et al., 2021; Wassen et al., 2021). Moreover,

∼2 m) concentrated the depositions in the vicinity of the trails sav-

in contrast to N, P is one of the least mobile mineral nutrients and

ing the rest of the area, but this then results in very high deposi-

legacies of P fertilization can last for centuries (Schelfhout et al.,

tion rates of 175 kg N ha−1 year−1 and 73 kg P ha−1 year−1 near

2017). This P immobility leads to difficult and slow on-site P removal

the paths. This N dose even nearly corresponds to the legal thresh-

management. Heavily fertilized, intensively managed agricultural

old set by the EU Nitrate directive (91/676/EEC) for N from livestock

grassland in Belgium yields about 14 mg dry matter ha−1 year−1

manure in the European Union. At this fertilization level, grasses dom-

and exports 52 kg P ha−1 year−1 (Cougnon et al., 2018). In Nardus

inate the vegetation and many forbs are outcompeted. A survey on

grasslands under restoration, however, removal rates are 2–20 kg P

French permanent grasslands, for example, showed that once N fertil-

ha−1 year−1 under mowing with hay removal (Schelfhout et al., 2019).

ization exceeds 150 kg N ha−1 year−1 , a presence of more than 10%

More drastic restoration techniques such as phytomining and topsoil

legumes in the biomass becomes very rare (Jeuffroy et al., 2015). The

removal can increase P exports, but are also more expensive and have

P dose of 73 kg P ha−1 year−1 largely exceeds the local legal threshold

strong impacts on other abiotic properties. Because the average P

for fertilization of agricultural grassland and arable land (i.e. 30–50 kg

fertilization by dogs in our study almost levels the annual export rates

P ha−1 year−1 depending on the P concentration of the soil) and the

by mowing with hay removal, it is clear that dogs can potentially have

potential P export through the grass harvest (see above). Leashing dogs

a strong impact on the vegetation and the management of these sites.

and removing their faeces reduced deposits to 77 kg N ha−1 year−1

Currently, these sites are under restoration management and mowing

and 2 kg P ha−1 year−1 in the vicinity of the path. The mowing fre-

with hay removal is applied to reduce P levels in the soil to promote

quency near the path could of course be enhanced (e.g. five to eight

plant biodiversity; this process will be significantly slowed down by

times per year) to export more nutrients than the rest of the area.

the import of canine P. In the forests, the current management of no

Third, enforcement also seems to have a clear effect when we com-

intervention and low-density livestock grazing is less oriented towards

pare data from site 1 where off-leash dogs only accounted for 8% of

P removal, but these management types presume a more or less closed

total dog numbers (a law enforcement officer strictly cautions and, on

P-cycle. Also, here, the effects of continuous P fertilization by dogs can

second infringement, fines every off-leash dog owner) with sites 2–4

eventually lead to eutrophication.

where the legal obligation to leash dogs is not enforced and off-leash dogs represented 29%–49% of dogs. [Correction added on 7 February 2022 after first online publication: percentages have been updated 9%

4.3 Recommendations for management: Applications

and 27%–52% to 8% and 29%–49%.] Obviously, enforcing codes does not always change behaviour (Oates et al., 2017). Fourth, specifically designed nearby fenced off-leash dog parks where dogs are allowed to

Given the potentially high fertilization rates by dogs in peri-urban

roam freely, together with a dog ban in sensitive oligotrophic ecosys-

ecosystems, guidelines for management should be directed towards

tems with plants adapted to nutrient-poor soils, could take away the

moderating these inputs such that critical load exceedance, biodiver-

pressure on areas that are important for biodiversity conservation.

sity loss and delay of restoration goals be avoided. Based on our

Finally, the hitherto often neglected fertilization effect by dogs should

results, we propose land managers, especially in ecosystems with

better be included in management plans, in media campaigns and in

species adapted to nutrient-poor soils, take actions to (i) stimulate vis-

public education programs with regard to dogs in (semi-)natural peri-

itors to take away solid faecal waste (the most important source of

urban ecosystems.

P) by emphasizing the fertilization effect of their dogs in addition to other more widely known negative impacts, for example on wildlife,

ACKNOWLEDGEMENTS

(ii) enforce leash use more stringently, (iii) establish more off-leash

We thank Dirk Reheul for comments on the manuscript, and Dirk

dog parks and (iv) consider more often entire dog bans in oligotrophic

Huyghe and Hanne Van Beneden for help with the dog counts in the

ecosystems. First, as faeces contained 97% of the P and 56% of the

Vinderhoutse Bossen and the Bourgoyen, respectively. We thank Dr.

N deposited, taking away the faeces using, for instance, disposal bags

Dejun Li, Dr. Shinichi Tatsumi and two anonymous reviewers for help-

and pooper-scooper stations can greatly decrease potential nutrient

ful comments that have considerably improved our manuscript.

8 of 9

DE FRENNE ET AL .

CONFLICT OF INTEREST The authors declare no conflict of interest. AUTHORS’ CONTRIBUTIONS All authors conceived the study design and co-wrote the paper. P.D.F., G.P.J.J. and P.V.G. collected and analysed data. DATA AVAILABILITY STATEMENT All raw data and code are available via figshare at https://doi.org/10. 6084/m9.figshare.17054171.v1 (De Frenne et al., 2021). ORCID Pieter De Frenne

https://orcid.org/0000-0002-8613-0943

REFERENCES Allen, J. A., Setala, H., & Kotze, D. J. (2020). Dog urine has acute impacts on soil chemistry in urban greenspaces. Frontiers in Ecology & Evolution, 8, 615979. https://doi.org/10.3389/fevo.2020.615979 Almería, S. (2013). Neospora caninum and wildlife. ISRN Parasitology, 2013, 947347. Atwal, J., Stockman, J., Gilham, M., Allaway, D., Renfrew, H., Bakke, A. M., & Watson, P. (2021). No observed adverse effects on health were detected in adult beagle dogs when fed a high-calcium diet for 40 weeks. Animals, 11, 1799. https://doi.org/10.3390/ani11061799 Arnesen, T. (1999). Vegetation dynamics following trampling in grassland and heathland in Selendet Nature Reserve, a boreal upland area in Central Norway. Nordic Journal of Botany, 19, 47–69. Beaver, B. V. (1999). Canine eliminative behavior. In B. V. G. Beaver (Ed.), Canine behavior: A guide for veterinarians (1st ed., pp. 267–287). W.B. Saunders Company. Beynen, A. C., Baas, J. C., Hoekemeijer, P. E., Kappert, H. J., Bakker, M. H., Koopman, J. P., & Lemmens, A. G. (2002). Faecal bacterial profile, nitrogen excretion and mineral absorption in healthy dogs fed supplemental oligofructose. Journal of Animal Physiology, 86, 298–305.

Bobbink, R., Hicks, K., Galloway, J., Spranger, T., Alkemade, R., Ashmore, M., Bustamante, M., Cinderby, S., Davidson, E., Dentener, F., Emmett, B., Erisman, J.-W., Fenn, M., Gilliam, F., Nordin, A., Pardo, L., & De Vries, W. (2010). Global assessment of nitrogen deposition effects on terrestrial plant diversity: A synthesis. Ecological Applications, 20(1), 30–59. Bonner, C., & Agnew, A. D. Q. (1983). Soil phosphorus as an indicator of canine faecal pollution in urban recreation areas. Environmental Pollution, 6, 145–156. Cargo-Froom, C. L., Fan, M. Z., Pfeuti, G., Pendlebury, C., & Shoveller, A. K. (2019). Apparent and true digestibility of macro and micro nutrients in adult maintenance dog foods containing either a majority of animal or vegetable proteins. Journal of Animal Science, 97, 1010–1019. Castrillo, C., Vicente, F., & Guada, J. A. (2001). Urinary energy losses in dogs fed commercial extruded foods. Journal of Animal Physiology and Animal Nutrition, 85, 237–241. Ceulemans, T., Stevens, C. J., Duchateau, L., Jacquemyn, H., Gowing, D. J. G., Merckx, R., Wallace, H., van Rooijen, N., Goethem, T., Bobbink, R., Dorland, E., Gaudnik, C., Alard, D., Corcket, E., Muller, S., Dise, N. B., Dupré, C., Diekmann, M., & Honnay, O. (2014). Soil phosphorus constrains biodiversity across European grasslands. Global Change Biology, 20(12), 3814– 3822. Cougnon, M., Vliegher, A.

D., Frenne, P. D., & Reheul, D. (2018). Phosphorus yield of forage grass species under intensive cutting management. Grassland Science in Europe, 23, 46–48. Decau, M. L., Simon, J. C., & Jacquet, A. (2003). Fate of urine nitrogen in three soils throughout a grazing season. Journal of Environmental Quality, 32, 1405–1413.

De Frenne, P., Cougnon, M., Janssens, G. P. J., & Vangansbeke, P. (2021). Data and code from: Nutrient fertilization by dogs in peri-urban ecosystems. figshare. https://doi.org/10.6084/m9.figshare.17054171.v1 de Molenaar, J. G., & Jonkers, D. A. (1993). De invloed van stikstof in de ontlasting van honden op de vegetatie in voedselarme bos- en natuurterreinen. IBN-DLO. De Schrijver, A., De Frenne, P., Ampoorter, E., Van Nevel, L., Demey, A., Wuyts, K., & Verheyen, K. (2011). Cumulative nitrogen input drives species loss in terrestrial ecosystems. Global Ecology and Biogeography, 20(6), 803–816. De Smet, B., Hesta, M., Seynaeve, M., Janssens, G., Vanrolleghem, P., & De Wilde, R. O. (1999). The influence of supplemental alpha-galactosidase and phytase in a vegetable ration for dogs on the digestibility of organic components and phytate phosphorus. Journal of Animal Physiology, 81, 1– 8. FEDIAF. (2019). European facts & figures 2019. The European Pet Food Industry. https://www.fediaf.org//who-we-are/european-statistics.html Fick, S. E., & Hijmans, R. J. (2017). WorldClim 2: New 1-km spatial resolution climate surfaces for global land areas. International Journal of Climatology, 37(12), 4302–4315. https://doi.org/10.1002/joc.5086 Flemish Environmental Agency. (2020). Results N deposition (in Dutch). https://www.vmm.be/data/verzuring-en-vermesting#section-2 Forster, G.

M., Hill, D., Gregory, G., Weishaar, K. M., Lana, S., Bauer, J. E., & Ryan, E. P. (2012). Effects of cooked navy bean powder on apparent total tract nutrient digestibility and safety in healthy adult dogs. Journal of Animal Science, 90, 2631–2638. Hautier, Y., Niklaus, P. A., & Hector, A. (2009). Competition for light causes plant biodiversity loss after eutrophication. Science, 324(5927), 636– 638. http://www.sciencemag.org/content/324/5927/636.abstract Hoogendoorn, C. J., Betteridge, K., Costall, D. A., & Ledgard, S. F. (2010). Nitrogen concentration in the urine of cattle, sheep and deer grazing a common ryegrass-cocksfoot-white clover pasture. New Zealand Journal of Agricultural Research, 53(3), 235–243. Jarvis, S. C. (2000). Soil-plant–animal interactions and impact on nitrogen and phosphorus cycling and recycling in grazed pasture. In G. Lemaire, J. Hodgson, A. Moraes, P. Cavarho, & C. Nabinger (Eds.), Grassland ecophysiology and grazing ecology (pp. 317–338). CABI International. Jeuffroy, M.-H., Biarnès, V., Cohan, J.-P., Corre-Hellou, G., Gastal, F., Jouffret, P., Justes, E., Landé, N., Louarn, G., Plantureux, S., Schneider, A., Thiebeau, P., Morison, M., & Vertès, F. (2015). Performances agronomiques et gestion des légumineuses dans les systèmes de productions végétales. In A. Schneider & C.

Huyghe (Eds.), Les légumineuses pour des systèmes agricoles et alimentaires durables (pp. 139–223). Editions Quae. http://prodinra. inra.fr/record/332853 Jones, L., Stevens, C., Rowe, E. C., Payne, R., Caporn, S. J. M., Evans, C. D., Field, C., & Dale, S. (2016). Can on-site management mitigate nitrogen deposition impacts in non-wooded habitats? Biological Conservation, 212, 464–475. https://doi.org/10.1016/j.biocon.2016.06.012 Lantinga, E. A., Keuning, J. A., Groenwold, J., & Deenen, P. J. A. G. (1987). Distribution of excreted nitrogen by grazing cattle and its effects on sward quality, herbage production and utilization. In H. G. Van Der Meer, R. J. Unwin, T. A. Van Dijk, & G. C. Ennik (Eds.), Animal manure on grassland and fodder crops. Fertilizer or waste? (pp. 103–117). Springer. Lenth, B. E., Knight, R. L., & Brennan, M. E. (2008). The effects of dogs on wildlife communities. Natural Areas Journal, 28, 218–227. McDowell, R. W., & Stewart, I. (2005). Phosphorus in fresh and dry dung of grazing dairy cattle, deer, and sheep: Sequential fraction and phosphorus-31 nuclear magnetic resonance analyses. Journal of Environmental Quality, 34, 598–607. Oates, S. C., Miller, M. A., Hardin, D., Dominik, C., Jessup, D., & Smith, W. A. (2017).

Daily relative dog abundance, fecal density, and loading rates on intensively and minimally managed dog-friendly beaches in central California. Marine Pollution Bulletin, 125(1–2), 451–458. https://doi.org/10. 1016/J.MARPOLBUL.2017.10.062

9 of 9

DE FRENNE ET AL .

Oelmann, Y., Broll, G., Hölzel, N., Kleinebecker, T., Vogel, A., & Schwartze, P. (2009). Nutrient impoverishment and limitation of productivity after 20 years of conservation management in wet grasslands of north-western Germany. Biological Conservation, 142, 2941–2948. Paradeis, B., Lovas, S., Aipperspach, A., Kazmierczak, A., Boche, M., He, Y., Corrigan, P., Chambers, K., Gao, Y., Norland, J., & DeSutter, T. (2013). Dog-park soils: Concentration and distribution of urine-borne constituents. Urban Ecosystems, 16(2), 351–365. https://doi.org/10.1007/ s11252-012-0264-1 Pegtel, D. M., Bakker, J. P., Verweij, G. L., & Fresco, L. F. M. (1996). N, K and P deficiency in chronosequential cut summer-dry grasslands on gley podzol after the cessation of fertilizer application. Plant & Soil, 178, 121–131. Perring, M. P., Manning, P., Hobbs, R. J., Lugo, A. E., Ramalho, C. E., & Standish, R. J. (2013). Novel urban ecosystems and ecosystem services. In R. J. Hobbs, E. S. Higgs, & C. M. Hall (Eds.), Novel ecosystems: Intervening in the new ecological world order (pp. 310–325). Wiley. https://doi.org/10.1002/ 9781118354186.ch38 Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., & R Core Team. (2021). nlme: Linear and nonlinear mixed effects models. R package version 3.1-153. https://CRAN.R-project.org/package=nlme Pinna, C., Vecchiato, C.

G., Bolduan, C., Grandi, M., Stefanelli, C., Windisch, W., Zaghini, G., & Biagi, G. (2018). Influence of dietary protein and fructooligosaccharides on fecal fermentative end-products, fecal bacterial populations and apparent total tract digestibility in dogs. BMC Veterinary Research, 14, 106. R Core Team. (2021). R: A language and environment for statistical computing (4.0.4). https://www.r-project.org Schelfhout, S., Mertens, J., Perring, M. P., Raman, M., Baeten, L., Demey, A., Reubens, B., Oosterlynck, S., Gibson-Roy, P., Verheyen, K., & De Schrijver, A. (2017). P-removal for restoration of Nardus grasslands on former agricultural land: Cutting traditions. Restoration Ecology, 25, S178– S187. Schelfhout, S., De Schrijver, A., De Bolle, S., De Gelder, L., Demey, A., Du Pré, T., De Neve, S., Haesaert, G., Verheyen, K., & Mertens, J. (2015). Phosphorus mining for ecological restoration on former agricultural land. Restoration Ecology, 23, 842–851. http://doi.org/10.1111/rec.12264 Schelfhout, S., De Schrijver, A., Vanhellemont, M., Vangansbeke, P., Wasof, S., Perring, M. P., Haesaert, G., Verheyen, K., & Mertens, J. (2019). Phytomining to re-establish phosphorus-poor soil conditions for nature restoration on former agricultural land. Plant & Soil, 440, 233– 246.

Schelfhout, S., Wasof, S., Mertens, J., Vanhellemont, M., Demey, A., Haegeman, A., DeCock, E., Moeneclaey, I., Vangansbeke, P., Viaene, N., Baeyen, S., De Sutter, N., Maes, M., van der Putten, W. H., Verheyen, K., & De Schrijver, A. (2021). Effects of bioavailable phosphorus and soil biota on typical Nardus grassland species in competition with fast-growing plant species. Ecological Indicators, 120, 106880. Sitters, J., Maechler, M. J., Edwards, P. J., Suter, W., & Venterink, H. O. (2014). Interactions between C:N:P stoichiometry and soil macrofauna control

dung decomposition of savanna herbivores. Functional Ecology, 28, 776– 786. Staude, I. R., Waller, D. M., Bernhardt-Römermann, M., Bjorkman, A. D., Brunet, J., De Frenne, P., Hédl, R., Jandt, U., Lenoir, J., Máliš, F., Verheyen, K., Wulf, M., Pereira, H. M., Vangansbeke, P., Ortmann-Ajkai, A., Pielech, R., Berki, I., Chudomelová, M., Decocq, G., . . . Baeten, L. (2020). Replacements of small- by large-ranged species scale to diversity loss in Europe’s temperate forest biome. Nature Ecology & Evolution, 4(6), 802–808. https://doi.org/10.1038/s41559-020-1176-8 Stevenson, A. E., Hynds, W. K., & Markwell, P. J. (2003). The relative effects of supplemental dietary calcium and oxalate on urine composition and calcium oxalate relative supersaturation in healthy adult dogs. Research in Veterinary Science, 75, 33–41. Van Dobben, H. F., Bobbink, R., Bal, D., & van Hinsberg, A. (2012). Overzicht van kritische depositiewaarden voor stikstof toegepast op habitattypen en leefgebieden van Natura 2000. Alterra. Wamelink, G. W. W., Goedhart, P. W., Roelofsen, H. D., Bobbink, R., Posch, M., & van Dobben, H. F. (2021). Relaties tussen de hoeveelheid stikstofdepositie en de kwaliteit van habitattypen (Rapport/Wageningen Environmental Research; No. 3089). Wageningen Environmental Research. https:// doi.org/10.18174/547752 Wassen, M. J., Schrader, J., van Dijk, J., & Eppinga, M. B. (2021).

Phosphorus fertilization is eradicating the niche of northern Eurasia’s threatened plant species. Nature Ecology & Evolution, 5(1), 67–73. https://doi.org/10. 1038/s41559-020-01323-w Weston, M. A., Fitzsimons, J. A., Wescott, G., Miller, K. K., Ekanayake, K. B., & Schneider, T. (2014). Bark in the park: A review of domestic dogs in parks. Environmental Management, 54, 373–382. Wickham, H. (2016). ggplot2: Elegant graphics for data analysis. SpringerVerlag. Wood, C. W., Cummins, K. A., Williams, C. C., & Wood, B. H. (2004). Impact of diet and age on element excretion from dogs. Communications in Soil Science and Plant Analysis, 35(9–10), 1263–1270. https://doi.org/10.1081/ CSS-120037544

SUPPORTING INFORMATION Additional supporting information may be found in the online version of the article at the publisher’s website.

How to cite this article: De Frenne, P., Cougnon, M., Janssens, G. P. J., & Vangansbeke, P. (2022). Nutrient fertilization by dogs in peri-urban ecosystems. Ecological Solutions and Evidence, 3, e12128. https://doi.org/10.1002/2688-8319.12128