Land sharing complements land sparing in the conservation of disturbance-dependent species

Alteration of natural disturbances in human-modified landscapes has resulted in many disturbance-dependent species becoming rare. Conservation of such species requires efforts to maintain or recreate disturbance regimes. We compared benefits of confining efforts to habitats in protected areas (a form of land sparing) versus integrating them with general management of production land (a form of land sharing), using two examples: fire in forests and grazing in semi-natural grasslands. We reviewed empirical studies from the temperate northern hemisphere assessing effects of disturbances in protected and non-protected areas, and compiled information from organisations governing and implementing disturbances in Sweden. We found advantages with protection of areas related to temporal continuity and quality of disturbances, but the spatial extent of disturbances is higher on production land. This suggests that an approach where land sparing is complemented with land sharing will be most effective for preservation of disturbance-dependent species in forests and semi-natural grasslands. Supplementary Information The online version contains supplementary material available at 10.1007/s13280-022-01820-1.

. The search terms used in the literature searches. Searches were performed in Scopus and using Google and Google Scholar, in English and Swedish. For search terms within brackets, separated by vertical bars, separate searches were performed for each of those terms and the rest of search terms in the search string, i.e. [grazing | mowing] AND biodiversity would mean that two separate searches were performed, one using the search string grazing AND biodiversity, the other mowing AND biodiversity.

Scopus
((fire* OR burn* OR wildfire*) AND (forest* OR woodland* OR wood*)) AND (fire-depend* OR fire-adapt* OR pyroph* OR saprox* OR dead-wood* OR deadwood* OR pyriscence OR heat-trigg* OR fire-trigg* OR wood-living OR woodfeeding OR fire-loving OR fire-favoured OR fire-favored OR woodbor* OR wood-bor*) AND (protect* OR natura2000 OR "natura 2000" OR "reserve" OR "reserves" OR "national park*" OR set-aside* OR certifi* OR "voluntary conservation" OR "sustainable forest*" OR "conservation forest*" OR ((strateg* OR priorit* OR design* OR plan* OR polic* OR management) AND conservation)) Google [protected | reserve | "national park" | natura2000] AND forest AND fire AND "fire dependent" OR "fire adapted" OR pyrophile OR pyrophilous OR saproxylic OR dead-wood

Swedish
Google skog AND brand AND brandgynnad OR pyrofil Google Scholar skog AND brand AND brandgynnad OR pyrofil

Semi-natural grassland
English Scopus (graz* OR mow* OR scyth* OR grassland* OR pasture* OR meadow* OR (harvest* AND (grassland* OR meadow*)) OR (cutting AND (grassland* OR meadow*))) AND ("grassland species*" OR indicator* OR specialist* OR "target species" OR "focal species*" OR management-depend* OR grassland-depend* OR grazingdepend* OR mowing-depend* OR (management w/5 depend*) OR (grassland* w/5 depend*) OR (grazing w/5 depend*) OR (mowing w/5 depend*) OR (disturbance* w/5 depend*)) AND (protect* OR "reserve" OR "reserves" OR natura2000 OR "natura 2000" OR "national park*" OR ((protect* OR "reserve" OR "reserves" OR natura2000 OR "natura 2000" OR "national park*") AND ((agri-environment* OR AES OR "common agricultural polic*" OR CAP OR ((agricultur* OR farm*) AND (subsidi* OR support))) AND (grassland* OR meadow* OR pasture*))) OR ((strateg* OR priorit* OR design* OR plan* OR polic*OR management) AND conservation)) Google [grazing | mowing] AND biodiversity AND grassland AND [protected | reserve | "national park" | natura2000] AND agri-environment OR "common agricultural policy" OR subsidies OR support S1.3 Questions related to semi-natural grasslands  How is management financed in protected and non-protected grasslands?  Are there any differences in the ease of getting AES payments for management between protected and non-protected grasslands?  Are there any practical differences between management of protected and non-protected semi-natural grasslands, e.g. regarding grazing animals, grazing intensity, mowing technique, timing and frequency, the ease of acquiring animals to graze or persons to perform mowing?  How detailed are the management plans for semi-natural grasslands receiving AESpayments? Do they specify, e.g., management method, intensity or timing? How common is it that payment and management for these grasslands is extended past the initial five or one-year period?  Are there any controls that check if protected and non-protected grasslands have been managed according to their management plan? How often do these occur? What happens if it is revealed that management has not occurred?  Are there any national or regional strategies available regarding occurrence of managed protected and non-protected grasslands across landscapes, to ensure, e.g., high connectivity between managed grasslands? To what extent are these strategies actually followed? There are multiple approaches for funding disturbances when adopting a land sparing strategy in both forests and semi-natural grasslands in Sweden. There is an annual budget for the general management of protected areas, set by the Swedish government and distributed by the Swedish Environmental Protection Agency (Swedish EPA) (Nilsson 2005), but EU LIFE programs (LifeTaiga 2021), and the Swedish framework for conservation of threatened species and habitat types can also be used for funding (Artdatabanken 2021). For land sharing strategies, subsidies for funding prescribed burning in non-protected forest certified by the Forest Stewardship Council (FSC) or Programme for the Endorsement of Forest Certification (PEFC) are available through the Swedish Forestry Agency (SFA) (Skogsstyrelsen 2021a, b), but according to personnel at major forestry companies, prescribed burning to fulfill the requirements of certified forests is usually funded by the companies themselves. In addition, agri-environmental scheme (AES) payments can be used to fund grazing and mowing in both protected and non-protected semi-natural grasslands (Jordbruksverket 2021), but according to personnel at the Swedish Board of Agriculture (SBA) there is no prioritisation of the allocation of AES payments based on protection status. However, if the aim or management plan of a protected grassland conflict with these requirements the opposite may be true. In fact, almost 40 % of seminatural grasslands in the Swedish Natura 2000-network have to use other funding sources than AES payments to finance management, with conflicts between AES requirements and the management plan of protected areas . Funding management of non-protected grasslands not receiving AES payments is even more difficult, and only about 50 % of this type of grasslands are managed each year (Glimskär et al. 2017).
Regardless of the funding source and protection status, there is variation in the funds available for maintaining disturbances. Personnel at the Swedish EPA and County Administration Boards (CABs) note that the annual budget for general management of protected areas can be highly variable. For example, in 2019, 860 million SEK were budgeted for conservation efforts such as management of protected areas, while in 2021 the budget was more than twice as large (Miljödepartementet 2019(Miljödepartementet , 2020. In contrast, they state that funding through LIFE projects can be more predictable, as the budget is fixed for the project duration, but that the subsidies available for prescribed burning in nonprotected forests can also be variable. According to personnel at SBA, the budget for AES payments for management of semi-natural grasslands is large enough that everyone eligible for payments receive them. Farmers receiving AES payments commit for management of semi-natural grasslands during a five-year period, and as long as requirements are met across this period and there is no change in the area of individual grasslands, the yearly payments stay the same across this period. However, the actual payments are not always large enough to cover the actual cost of grazing or mowing in individual grasslands, as the amount payed is based on the average cost of grazing or mowing of one ha grassland (Jordbruksverket 2021). In addition, the EU Common Agricultural Policy, which govern AES is reformed at regular intervals (European Union 2021), making it more difficult to predict future compensation levels for management of semi-natural grasslands. Personnel at CABs state that because of the variation in available funding it is more difficult to predict how much will be available for maintaining disturbances in forests and semi-natural grasslands each year, or plan disturbances in the long-term. Some years, there is a risk that not all necessary disturbances can be financed, particularly in protected areas. In these cases, CABs prioritise disturbances of areas requiring annual disturbances (e.g. semi-natural grasslands), rather than areas that are less sensitive to the exact timing of disturbances (e.g. prescribed burning in forests).
S3.2 Differences in controls of the occurrence of disturbances with land sparing and land sharing strategies There are no guidelines of how often protected forests and semi-natural grasslands should be visited to control whether the disturbances detailed in the management plan actually occurs, but some monitoring of protected areas e.g. focus on the total area burned. In addition, management plans for protected areas often specify that the occurrence of disturbances should be checked. However, the exact method and frequency for controls is generally not specified, and there are no official statistics on how common discrepancies are. Personnel at CABs state that protected areas with higher disturbance frequency and more visitors are generally checked more often, and that if disturbances have not occurred, a solution is found to ensure the occurrence of future disturbances. In addition, personnel at CABs also state that if CABs manage disturbances internally, it is more likely that disturbances occur as they should.
There is an annual audit by FSC/PEFC-certified forest owners. According to personnel at FSC, during this audit the certifier gather documentation from individual forest owners to verify that the registered forest area burned fulfill the certification requirement. If this is not the case, and cannot be justified by, e.g., security reasons (e.g. burning has been impossible due to weather conditions), this has to be remedied. If the total burned area is still too small at the next audit, the forest owner can no longer sell certified timber, and they may have to be re-certified. However, there are no official statistics on how common discrepancies are. One county in northern Sweden (Västerbotten) compile a database of all fires occurring in both protected and non-protected areas within the county, which can be used to control whether or not planned forest fires occur.
According to personnel at SBA and CABs, about 5 % of all semi-natural grasslands managed using AES payments are visited each year, to ensure all requirements are fulfilled. Which grasslands are visited is mainly chosen at random. This means that most visited grasslands are non-protected, and that for a large proportion of grasslands receiving AES payments it is never verified whether management actually occurs. The most common discrepancy is that grazing or mowing have not occurred (yet), or that the management intensity is too low. If this is not rectified in time, this can lead to a deduction of payments, and if management has not occurred for years, AES payments can be cancelled entirely. Several studies have assessed the management in semi-natural grasslands, most of which are likely receiving AES payments, and while some reveal that most surveyed grasslands have sufficient management (Glimskär et al. 2017), others reveal that for many of the surveyed grasslands management was insufficient or lacking (Hägglund et al. 2017;Nordberg 2013;Spörndly and Glimskär 2018;Stenström 2019Stenström , 2020. Up to 50 % of semi-natural grasslands without AES payments were not managed (Glimskär et al. 2017).

S3.3 Differences in monitoring of the effects of disturbances with land sparing and land sharing strategies
The Swedish EPA have developed general guidelines for monitoring of protected areas, but also specific manuals for monitoring of individual habitat types. Monitoring of protected areas is financed by the budget for the general management of protected areas, and it is generally recommended that this monitoring is repeated across 12 year intervals (Haglund 2010). In protected forests, monitoring focus on remote sensing to estimate the area of different forest types, but in some forest types the occurrence of fire traces such as burned trees is also monitored. In protected areas with fire prescribed in the management plan, monitoring directly following a natural or prescribed fire is recommended to estimate the total burned area and the total amount of burned and dead wood. It is also recommended that monitoring should be repeated at regular intervals, e.g. to investigate the occurrence of pyrophilous species (Haglund 2010;Kellner 2012). In semi-natural grasslands, monitoring focus on several variables such as aerial image analysis of the cover of shrubs and trees, the vegetation height at the end of the growing season (to ensure annual management actually occurs), and the occurrence of grassland plant, bird and insect species (Haglund 2010;Haglund and Vik 2010). The data from the monitoring is maintained by CABs. Every sixth year, the conservation status of threatened habitats and species listed in the EU Habitats Directive, which are often protected in Natura 2000-areas, has to be reported (Naturvårdsverket 2020). Thus, this also contributes to monitoring of disturbance-dependent habitats and species in protected areas.
According to personnel at forestry companies, monitoring after prescribed fire in FSC/PEFC-certified forests mainly focus on the total area burnt rather than any conservation benefits. However, the forestry companies state that their aim is to standardise and improve the monitoring to also focus on factors such as the amount of burned and dead wood and the occurrence of pyrophilous species. There is no formal monitoring specifically of non-protected semi-natural grasslands (with or without AES payments). However, there is a monitoring program surveying the occurrence of bumblebees, butterflies and plans in approx. 700 semi-natural grasslands across Sweden (Sveriges Lantbruksuniversitet 2021). While this monitoring occurs in both protected and non-protected grasslands with and without AES payments, this data can still provide some general conclusions on the effects of disturbances in semi-natural grasslands.

S3.4 Differences in spatial continuity of disturbances with land sparing and land sharing strategies
Neither management plans for protected areas, the requirements for FSC/PEFC-certified forests nor requirements for semi-natural grasslands managed using AES payments consider the occurrence of disturbances on landscape scales (Forest Stewardship Council 2020; Jordbruksverket 2021). Nevertheless, various strategy documents available from the Swedish EPA, SFA, SBA, CABs and forestry companies do emphasise the importance of planning disturbances across landscapes.
National strategies emphasise the importance of identifying landscapes with high conservation values to guide and prioritise conservation efforts such as prescribed burning and protection of areas (Naturvårdsverket 2013;Naturvårdsverket Skogsstyrelsen 2017;Nitare 2014). There are also strategies related specifically to the occurrence of fire, developed by individual CABs. These specify in which protected areas future prescribed burnings should occur, based on the location and fire history of an area, and the occurrence of pyrophilous species. Often, one aim of these strategies is to ensure a continuous supply of recently burned areas within landscapes (e.g. Berglund 2012;Lindhagen 2009). This topic is sometimes also considered in strategies related to green infrastructure (e.g. Länsstyrelsen Örebro län 2019). Forest owners with FSC/PEFC-certified forests are required to establish a landscape plan of their forest holdings, detailing forest types and age, location of protected areas and areas appropriate for conservation efforts such as prescribed burning. This information can then be used to plan which areas should be burned to fulfill the area requirement (Forest Stewardship Council 2020). Eco-parks are large areas (at least 1 000 ha) owned by forestry companies, within which they combine regular forestry activities with conservation efforts. These parks have a management plan, and in at least some cases these plans specify the location of future prescribed fires within the parks (Sveaskog 2021). It is unclear how well available regional strategies and landscape plans are actually followed. Personnel at both CABs and forestry companies state that they try to follow them when determining in which areas to perform prescribed burning, but that this is not always possible for security reasons or because areas are too small to burn efficiently without risking fire spread. Still, even if these strategies are not always followed exactly, CABs and forestry companies try to ensure that the occurrence of prescribed burning is spread throughout landscapes. However, individual strategies and plans do not encompass the entirety of the landscape, as these only consider the forests owned by the agency or company responsible for establishing them, and not all forest in the landscape.
In their strategy for preserving biodiversity in the agricultural landscape, SBA emphasise the importance of increasing the area of semi-natural grassland in landscapes, creating and improving dispersal pathways between grasslands, and the importance of preserving valuable grasslands (Wallander et al. 2019). Green infrastructure plans, created by all CABs, also focus on semi-natural grasslands in a wider landscape perspective. These plans identify landscapes with high total area of semi-natural grasslands and high connectivity between grasslands, but also identify where connectivity can or should be improved, through appropriate management of road verges, power line corridors or grasslands fragments (Alsén and Kruys 2019; Berlin and Niss 2019; Länsstyrelsen Örebro län 2019). For all these strategies and plans, it is unclear how well they are followed.

S3.5 Differences in the method and intensity of disturbances between land sparing and land sharing strategies in forests
The total area burned during prescribed fires is generally small, regardless of protection status (average size is 12 ha). However, the forest type burned sometimes differ depending on protection status, which will affect the conservation benefits of a fire: in protected areas, mature and unmanaged forests are generally burned, while 67 % of prescribed burning performed by forestry companies occur on clear-cuts (Ramberg et al. 2018). This type of fire will have limited benefits for pyrophilous and saproxylic species, as there will only be limited availability of burned and dead wood after fire, and as new forests will quickly be planted on burned areas (Heikkala et al. 2017;Heikkala et al. 2016;Hyvärinen et al. 2009;Ranius et al. 2014;Toivanen and Kotiaho 2007). According to the Swedish EPA, only a minority of burnings performed by forestry companies can be classified as conservation burning (Wikars 2006). However, burning on clear-cuts can still have positive effects on some pyrophilous vascular plants and ground-dwelling fungi that require high temperatures for reproduction (Nilsson 2005).
Salvage logging consists of logging of trees that are dead or dying as a result of, e.g. a forests fire or storm felling, to decrease economic losses or prevent outbreaks of pest species such as the spruce bark beetle (Ips typographus) (Lindenmayer et al. 2008). Thus, if salvage logging occurs after a natural or prescribed forest fire, this will limit the benefits to pyrophilous and saproxylic species (Cobb et al. 2010;Cobb et al. 2011;Koivula and Spence 2006;Kotliar et al. 2002;Lindenmayer and Noss 2006;Thorn et al. 2018), but can still have positive effects on organisms benefitting in open forests, like flowering plants and pollinators (Heil and Burkle 2018). According to personnel at CABs, salvage logging never occurs in protected areas after a fire, unless there is a great risk of pest species outbreaks that could threaten the conservation values in a protected area (Naturvårdsverket 2021). The requirements for FSC/PEFC-certified forest does not specify that salvage logging is forbidden after a fire. However, if enough burned or dead wood is left after a fire (at least 15 % of the tree volume remaining after fire), the net burned area can be multiplied by an upward adjustment factor (Forest Stewardship Council 2020; Programme for the Endorsement of Forest Certification 2017), meaning less area has to be burned to fulfill the certification requirements. Thus, this can create an incentive to avoid salvage logging. However, regulations for production forests limits the amount of dead wood allowed per ha after, e.g., storm felling or fire (maximum 5 m3 conifers per ha) (Skogsstyrelsen 2020). According to personnel at forest companies it varies how common salvage logging is after fire in FSC/PEFC-certified forests: while there is generally no salvage logging after fire in voluntary setasides, logging sometimes occur in non-set-asides. To what extent salvage logging occurs after natural forest fires depends on the protection status and the size of the fire. After a mega-fire, affecting 13 100 ha of mostly production forest in central Sweden in 2014, about 40 % of the burned area was salvage logged (with most of the rest of the area being protected as a result of the fire) (Gustafsson et al. 2019).
According to personnel at CABs and forestry companies, it is generally challenging to perform prescribed burning regardless of protection status, due to the high security requirements. Prescribed burning in both protected and non-protected forests can be performed by CABs or forestry companies, but most often specific companies are hired. There can be some differences in the flexibility in how and when prescribed burning is performed depending on protection status. For example, according to personnel at forestry companies, flexibility can be higher in non-protected forests, as there is no set management plan that needs to be followed. Fire severity is a factor that affects, e.g., how much of individual trees are burnt and the total number of trees burned, and can affect the response of species and ecosystem processes (Keeley 2009). It is unknown if there are any general differences in the severity of prescribed fires in protected and non-protected forests. However, an assessment from northern Sweden revealed that prescribed fires generally have higher conservation benefits compared to natural forest fires (e.g. as a consequence of higher burn severity) (Sundin 2006).