A need for incentivizing field hydrology, especially in an era of open data: discussion of “The role of experimental work in hydrological sciences – insights from a community survey”*

ABSTRACT

The sharing of data and collection of new data are both essential, but they are not inherently complementary. When data are openly available, researchers may be motivated to use those data rather than collect more because field work has costs and risks. The competitive advantage to those who do not put resources towards fieldwork may discourage field hydrology. Allocating efforts towards generating field data, which benefits hydrological sciences, is not necessarily best for individual hydrologists, especially in an era of open data. The objective of this work is to open a conversation on whether individuals’ best interests may contrast with the community’s desire for new observations. If the community wants new field observations, there is a need to consider the shifting balance of incentives and disincentives for pursuing field studies in hydrology.

KEYWORDS:

• fieldwork
• open-access data
• data repositories
• social dilemma
• experimental hydrology

Editor A. Castellarin; Associate editor not assigned

Hydrologists praise fieldwork. Results from a community survey by Blume et al. (2017) indicate that advancements in hydrological sciences are expected to be primarily dependent on new observations. Therefore, the community asks for more fieldwork, new field measurements, and improved integration of field efforts with modelling (Blume et al. 2017). Despite this desire for field data, hydrological research appears to be increasingly dominated by modelling studies (e.g. Burt and McDonnell 2015, Vidon 2015). If field hydrology is declining, this is fundamentally problematic because the advancement of hydrological knowledge is ultimately dependent on observations (Kirchner 2006). Not only would a decline in field hydrology result in less field data, it could also result in field knowledge and skills atrophying as future hydrologists receive less field training (Vidon 2015).

Reasons to not conduct fieldwork

This (perceived) shortage of field hydrology is not surprising because the individuals that allocate effort towards collecting data are often disadvantaged: fieldwork comes at a cost of additional time, money, and risk, which is considered to impede publication efficiency (Blume et al. 2017). We expect that this risk especially discourages establishing new sites, particularly where workforce, funding, or infrastructure are lacking. While the hydrology community may mostly agree that more field data allows hydrology to best progress (Blume et al. 2017), the community is composed of individual hydrologists with individual incentives. A prominent individual incentive (in science) is building one’s own career, which is facilitated by publishing many papers, especially early on (Hirsch 2005, Abbott et al. 2010, Aitkenhead 2013, Geman and Geman 2016). By using already-collected data, the risks and costs of fieldwork can be avoided. In addition, data syntheses can tackle questions of greater scope and greater scale than case studies, because they allow us to better understand the generality of observations (Sivapalan and Blöschl 2017). Therefore, there are advantages to focusing on analysing others’ data rather than conducting fieldwork, especially for early career hydrologists who have greater publication pressures. Given these advantages of data re-use, we argue that there are often inadequate incentives for individuals to conduct fieldwork. This situation is at odds with the community’s desire for the collection of more field observations (Blume et al. 2017).

Benefits and costs of sharing

There is a near-unanimous opinion among hydrologists that all data should be shared (Blume et al. 2017). However, sharing affects the tradeoffs between doing and not doing fieldwork. Researchers will use data because they are available, even if they are not ideal. For example, studies of rainfall–runoff responses have largely relied on existing stream and precipitation gauge networks, designed for operational purposes, even though they may have limited usefulness for hydrological research (Kirchner 2006). As more primary data from hydrological research become shared, hydrologists (whether field hydrologists, modellers, or meta-analysers) can conduct better research because more and better data will be available for re-use (Sivapalan and Blöschl 2017). Data collected specifically for hydrological research can be highly valuable because they are likely to be aimed at supporting new hydrological understanding and hypothesis testing. So, when data collected by field hydrologists are made open, modellers benefit from access to those data and the collectors lose the advantage of exclusive access to those data. We believe the community’s desire for more data warrants proactive discussion of how to balance these factors that we think are influencing the prevalence of field hydrology.

A common good versus individual benefit game

The dichotomy of contrasting individual incentives and community needs in hydrological data is a social dilemma: optimal collective benefits are inconsistent with individuals’ best strategies. One well-known example of a social dilemma is what Hardin (1968) coined as “The Tragedy of the Commons”: if there is a freely available public resource, it is in the individual’s best interest to use that resource, even if this results in degradation of the resource and prevents the best outcome for the collective (Lloyd 1833, Dawes 1973). The situation of field hydrology differs from the tragedy of the commons because use of data does not degrade the common resource. Nevertheless, the availability of open data encourages individuals to use available resources, rather than to generate more. A decreased focus on generating data likely represents a suboptimal outcome because the community sees more data as a pathway for hydrology to best progress (Blume et al. 2017).

Experimental games have been used to better understand participant behaviour in social dilemma scenarios (Hauert et al. 2002, Santos et al. 2008). For example, “public goods” games involve players having the option to contribute money to a collective pot; the amount in the pot is multiplied and then evenly distributed back to the participants. A higher proportion of contributors results in an increased overall reward. However, any individual benefits most by contributing nothing (so-called “free riding”; Andreoni 1988), so this is the rational strategy for any player. Empirically these games have shown that individuals typically “free ride” rather than contribute to the public good when the game is iterated (Andreoni 1988). In hydrology, collecting and sharing data is like contributing money to the public pot. Of course, free riding is not bad because data re-users advance our understanding by using extant data. Yet, some must contribute for there to be collective net benefits. One important finding from these games is that implementing incentives can result in greater collective benefits (e.g. Pronk et al. 2015).

The moral imperative of field hydrology is not enough

The community should not expect individuals to act against their own interests to solve a field data shortage. Past commentaries have asked early career scientists to do more fieldwork (e.g. Vidon 2015). However, in the highly competitive academic landscape, the community should not expect those that are not tenured to undermine their own career prospects by pursuing the riskier or less efficient routes. Hardin (1968) warns against such moral imperatives: “Responsibility is a verbal counterfeit for a substantial quid pro quo. It is an attempt to get something for nothing”. Hardin’s pessimistic comment is not universally applicable because some individuals may act altruistically or champion cultural and institutional changes that benefit field hydrologists. Ultimately, established community members play a crucial role in who is funded, hired, and promoted. Indeed, public goods experiments empirically demonstrate that some individuals will always act altruistically (Andreoni 1988), and individuals having a reputation for altruism can inspire others to act similarly (Milinski et al. 2002). While cultural changes can be effective in solving social dilemmas (Ostrom et al. 1999), deliberate and active incentives for fieldwork are worth consideration.

Privatized data is not the solution

A market could exist where users pay data generators to access field data. This would incentivize field hydrology by compensating the monetary costs of fieldwork. However, such a system is at odds with the community’s desire for open data (Blume et al. 2017) and it could enable data tycoons, create severe entry barriers, impede reproducibility, and keep data out of the hands of those who might best advance knowledge. Furthermore, exploratory modelling work might decline because up-front costs could dissuade modellers from experimenting with data. Similarly, field hydrologists may forgo bold experiments and instead opt to collect data that are known to have market value to modellers and data re-users. Thus, a money-based data market has unacceptably regressive consequences.

Rewards incentivize

Some of the disadvantages of a money-based data market could be avoided in a market where co-authorship is the form of payment. This system reduces barriers set up by financial inequalities across the global science community. A likely benefit of including primary data collectors on projects is better science because of their understanding of the data and sites. However, such interactions may never arise if data are simply accessed from public repositories. The concept of offering co-authorship for data has been adopted by networks such as Ameriflux (Guy et al. 2007) or Sapfluxnet (Poyatos et al. 2016), and it encourages contribution of high-quality data and meta-data. However, individuals and smaller fieldwork groups with unique datasets may be less able to leverage their data compared to networks that can exercise a form of collective bargaining. Although it is difficult to assess what level of data contribution constitutes co-authorship, offering co-authorship is usually beneficial (Budker and Kimball 2016). Regardless of the pros and cons of co-authorship-for-data exchanges, dataset licensing and the growing presence of data papers with DOIs also provide improved valuation of fieldwork and data management (Wilkinson et al. 2016, Goldstein et al. 2017). Consistent formal rewards for field hydrologists will help to incentivize the collection and sharing of field data.

Data openness is less discouraging for field hydrologists if the costs and risks of fieldwork are mitigated. For example, funding agencies can increase the fraction of money going towards new observations. Grants do not only provide money for research, but are also important for securing one’s career. An alternative approach of sharing risk could also be achieved, for example by collectively pooling small individual contributions of time and money to execute grand field experiments with shared data. Such initiatives may generate bold novel field data, which disproportionately advances science (Burt and McDonnell 2015, Tauro et al. 2018). Citizen science is another means of creating novel hydrological data (Buytaert et al. 2014), because many volunteers can monitor more sites than a typical research team can. These are just a few ideas of how to incentivize field hydrology.

Navigating forward

The community’s call for more field observations (Blume et al. 2017) diagnoses a problem, but solutions cannot follow until the underlying causes of the problem are also understood. We argue that the advantages of using data over collecting data are a major underlying cause. While we agree that the openness of data is crucial, the community should also consider that any apparent declines in field hydrology may stem from its relative disadvantages compared to synthesis work; these disadvantages might become more pronounced in this era of open data. We do not negate that field hydrologists also have many advantages. Nevertheless, solutions are needed because this inherent dichotomy between individual and collective goals is unlikely to passively subside. We have outlined a diversity of potential incentives, many of which are already in some form of implementation. A broad, adaptive approach is likely necessary because individuals differ in how they respond to incentives; for example, the factors that influence whether individuals contribute data to repositories depend on career stage, personality, and institutional environment (Linek et al. 2017). If the community wants more field observations, it is worth considering what factors influence the decision to do field hydrology. Otherwise, field hydrology may decline, no matter what praise it receives.

Acknowledgements

We thank Theresa Blume, Ilja van Meerveld and Markus Weiler for publishing the survey that provided impetus to writing this manuscript. Theresa Blume and Salvatore Grimaldi provided helpful reviews. Ilja van Meerveld and James Kirchner offered comments that guided the writing of this manuscript. Josh Larsen and Greg Goldsmith provided comments on an earlier draft.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

*Blume, T., Van Meerveld, I., and Weiler, M., 2017. The role of experimental work in hydrological sciences – insights from a community survey. Hydrological Sciences Journal, 62 (3), 334–337. doi: 10.1080/02626667.2016.1230675