The industrial fixation of nitrogen gas (N₂) to ammonia (NH₃), the Haber–Bosch (HB) process

is considered to be one of the most important chemical reactions in the world (e.g., Smil 2001) and ‘the greatest single experiment in global geo-engineering ever made’ (Sutton et al 2013, p. 4). This reaction has produced the nitrogen (N) fertilizers that have contributed to the ‘green revolution’ responsible for increased food production that has supported the expansion of human population from ∼2 billion in the early 20th century to >7 billion people today (Smil 1999, Erisman et al 2008).

Prior to World War II, the creation of reactive N was largely due to natural processes, including biological N fixation and lightning, and population expansion kept pace with its creation (Galloway et al 2002). After the mid-1940s and the commercialization and scaling up of the HB process, the manufacture and use of N expanded rapidly, from <10 MT N yr⁻¹ in 1950 to >170 MT N yr⁻¹ in 2013 (MT = megatonnes; figure 1(A); Constant and Sheldrick 1992, FAO 2012, Heffer and Prud’homme 2013). In fact, 85% of all synthetic N fertilizers have been created since 1985 (Howarth 2008).

Zoom In Zoom Out Reset image size

The rate of change in use of N fertilizers has eclipsed that of phosphorus (P) fertilizers in large part due to this large-scale capacity for anthropogenic synthesis. Global use of N fertilizer has increased nine-fold, while that of P has increased three-fold (Sutton et al 2013; figure 1(B)) and the upward trends in N : P in fertilizer application (figure 1(C)) are apparent in virtually all regions of the globe (figures 1(D)–(F)). In the US, it has been estimated that there has been at least a five-fold increase in reactive N use on average compared to pre-industrial time (Houlton et al 2013), but this increase is spatially variable ranging from negligible to 35-fold in different areas (Sobota et al 2013).

Although the HB process is the conversion of N₂ to NH₃, nearly 60% of all N fertilizer now used throughout most of the world is in the form of urea (CO(NH₂)₂; Constant and Sheldrick 1992, Glibert et al 2006, IFA 2014; figure 1(A). Urea production is an extension of the HB process, being produced by reacting CO₂ with anhydrous NH₃ under pressure at high temperatures. World use of urea as a fertilizer and feed additive has increased more than 100-fold in the past four decades (Glibert et al 2006). From 2001 to 2010, global urea use grew on average at a rate of 3.8% yr⁻¹, and it is projected that from 2012 to 2017 an estimated 55 new urea manufacturing plants will be constructed worldwide, with half of these located in China (Heffer and Prud’homme 2013), contributing to another anticipated doubling by 2050 (Glibert et al 2006). Multiple factors, including the less explosive nature of urea relative to ammonium-nitrate fertilizer (NH₄NO₃), make the transportation and storage of this synthetically produced N form much safer and the preferred choice for agricultural applications.

The incorporation of agricultural N into plant biomass is extremely inefficient. Although the efficiency of N use in experimental fields may be much higher than the global average of 50% (Balasubramanian et al 2004), under practical conditions it is difficult to equate the N supply from fertilizer and from soil organic matter mineralization with the dynamics of crop N uptake demand (Dobermann and Cassman 2005). Considering the complete food chain, only ∼10–30% of N applied actually reaches human consumers (Galloway et al 2002, Houlton et al 2013). The difficulty in improving N use efficiency in agriculture lies in the high mobility of N in the soil-plant system, and the variety of potential loss pathways ranging from NH₃ volatilization, denitrification, leaching and runoff, and other N transformation processes (Bouwman et al 2009). Urea inputs are typically hydrolyzed to NH₄⁺ in soil, but losses via volatilization and from runoff can be large and depend on the timing of application, weather, soil temperature and pH and other factors (Khakural and Alva 1995, Wali et al 2003). In regions such as China where the rate of fertilization has risen rapidly, the rate of soil retention of the excess N is actually declining (Cui et al 2013), leading to further environmental leakage.

The recovery of fertilizer P in crop products is also low (Syers et al 2008) but its biogeochemistry leads to proportionately greater retention within soils and sediments than N (e.g., Rhue and Harris 1999, Smil 2000, Bouwman et al 2009). The accumulation of residual soil P due to large fertilizer P surpluses over crop uptake during the 1960s, 1970s, and 1980s has led to an increased pool of plant available P in soils of most industrialized countries; a similar development was seen in later decades in India and China (Sattari et al 2012, 2014).

Among the various fates of the ‘leaked N’ are pathways that ultimately lead to N enrichment of lakes, rivers, and coastal waters. The major pathways for this leaked N include direct runoff, estimated to range up to 40% of inputs in large rivers (e.g., Howarth et al 2006) and atmospheric volatilization of NH₄⁺, and together these pathways can comprise more than half of the N input (e.g., Galloway et al 2004). P also runs off to receiving waters, but given the aforementioned change in patterns of fertilizer use and its biogeochemistry, the stoichiometry of the runoff has also changed in the last decades, leading to increasing N : P in receiving waters (e.g., Glibert et al 2013). It has been estimated that the atmospheric deposition of nutrients in the ocean is now ∼20 times the Redfield ratio for N : P (Jickells 2006, Peñuelas et al 2012) and these changes are also having large consequences for N : P stoichiometry in lakes (Elser et al 2009). This change in stoichiometry has been further compounded since the mid-1980s and 1990s when the major industrialized nations began curtailing P use by removing it from detergents and by upgrading sewage treatment processes which generally are more efficient in removing P than N (Litke 1999, Van Drecht et al 2009).

In general, urea concentrations in aquatic ecosystems are less than those of the inorganic N forms NO₃⁻ and NH₄⁺, but depending on proximity of the land source to the water body, concentrations of urea in lacustrine, estuarine and coastal waters may be high, particularly when runoff occurs from heavily fertilized areas (Glibert et al 2005, 2006). Concentrations up to 25–50 μM N have been reported in tributaries of the Chesapeake Bay (Lomas et al 2002, Glibert et al 2005), and nearshore waters adjacent to the heavily fertilized Yaqui Valley, Mexico (Glibert et al 2006), among other coastal areas (Kudela et al 2008, Switzer 2008). Urea concentrations vary from undetectable to 150 μM N in Lake Kinneret, Israel (Berman 1974), in Polish lakes (Siuda and Chrost 2006) and in lakes of Central Canada where these concentrations also represented 10–50% of bioavailable N (Bogard et al 2012). Urea is also part of the dissolved organic N pool that is typically a very large component of the available N in nutrient rich waters (Glibert et al 2006, Solomon et al 2010).

Eutrophication is the process by which waters are enriched in nutrients, leading to effects such as increased algal growth and development of high biomass blooms, changes in species diversity of both the primary and secondary producers, reductions in dissolved oxygen, fish kills, and the increased frequency of harmful algal blooms (HABs; Nixon 1995, Cloern 2001). HABs are those proliferations of algae that can cause ecological harm to food webs when they accumulate in massive quantities, and they can cause ecological, human health, and economic impacts when these cells produce toxic or other bioactive compounds and when the decay of high algal biomass results in hypoxia (Hallegraeff 1993, Glibert et al 2005, Backer and McGillicuddy 2006). The most common HABs are either dinoflagellates or cyanobacteria, although not all dinoflagellates or cyanobacteria are harmful, and not all HABs are made up of these species groups. Cyanobacteria are the HAB functional group of proportionately greater concern in freshwater, while dinoflagellates are the HAB functional group of greater concern in estuarine and marine waters. HABs have been expanding globally, in spatial extent, in duration of blooms, and in intensity (Anderson et al 2002, 2008, Glibert et al 2005) and a critical question has been the extent to which this change is associated with eutrophication and/or accelerated by climate or other factors (e.g., Paerl and Scott 2010). There are many reports of increases in HABs associated with eutrophication or nutrient loading (e.g., Anderson et al 2002, 2008, Glibert et al 2006, 2010, Heisler et al 2008), but the complexity of the relationship is far from understood. Many nutrient reduction strategies are focused on reducing algal biomass (e.g., total chlorophyll), while appropriate efforts or best management practices that may specifically prevent toxic HABs have remained much more challenging (Jewett et al 2008).

One of the most central tenets of aquatic science is that algal biomass and production in lakes and other freshwaters is limited by the availability of P, while that in marine waters is more often limited by the availability of N (e.g., Ryther and Dunstan 1971, Schindler 1977). However, P limitation in lakes is not universal (Lewis and Wurtsbaugh 2008) as some regions are either naturally (Finlay et al 2010) or culturally enriched (Bennett et al 2001) in P relative to N, and excess N loading is changing the nutrient stoichiometry and limiting element in some coastal areas (e.g., Sylvan et al 2006). It is commonly assumed that to control eutrophication the only focus should be on that nutrient which is classically considered ‘limiting.’ Such an argument is typically extended to promote enhanced P control over N control (e.g., Schindler et al 2008, Wang and Wang 2009) for multiple reasons, among which it is often assumed that N₂-fixing toxic cyanobacteria will be favored when N is limiting. Contrary to this common perception, in enriched systems, some toxic cyanobacteria are more common under elevated N : P conditions or when total N (TN) concentrations are high, and several cross comparative studies show that total cyanobacterial biomass can be predicted from increasing TN concentrations and from changes in TN to total P ratios (Smith 1983, Downing et al 2001, Kosten et al 2012, Dolman et al 2012). Therefore, species composition is important to consider (Dolman et al 2012) as, in fact, some of the most toxic cyanobacteria, namely Microcystis sp., are not N₂-fixing species.

Further arguments for P relative to N control are that P does not have a gaseous form and therefore cannot be permanently removed from lakes, whereas N can be lost to the atmosphere through denitrification provided conditions are optimal to favor coupled nitrification–denitrification (Seitzinger 1988, Cornwell et al 1999). However, rates of N₂-fixation do not necessarily offset N limitation (Scott and McCarthy 2010, Lewis et al 2011) and not all N enrichment is removed through denitrification resulting in eutrophication related problems that may be spatially and temporally displaced from the nutrient source (Conley et al 2009, Pearl 2009, Glibert et al 2011).

Here we make the case that accelerated P reduction over N control is insufficient in reducing the frequency of HABs; we base this on our growing understanding of HAB physiology and we support this with long-term trend data from one of the most eutrophic regions of the world, China. We also argue that increasing N : P environments further favor HABs when the N form is disproportionately in chemically reduced form (i.e. urea, NH₄⁺) relative to chemically oxidized form (i.e., NO₃⁻). We thus propose that the HB-HAB link is a function of (1) the inefficiency of incorporation of N fertilizers in the food supply chain, the leakiness of the N cycle from crop to table, and the fate of the lost N to the environment as described above; and (2) adaptive physiology of many HABs to thrive in environments in which there is excess N relative to classic nutrient stoichiometric proportions and where chemically reduced forms of N are increasing. We review physiological evidence for this at the cellular level, and we highlight long-term trends in HABs and fertilizer in China demonstrating increased HAB proliferations for systems in which N : P stoichiometry has been altered due to continued N loading with P reduction and the greater proportional use of urea as the main N fertilizer.

There are a number of specific physiological strategies that allow certain types of algae to thrive under conditions of elevated N : P availability relative to classic Redfieldian proportions (Redfield 1934), but not all cells necessarily have all such adaptive strategies (Glibert and Burkholder 2011). The first strategy is a low overall requirement for P. Very small cells, such as picocyanobacteria, have a lower requirement for P due to the smaller need for structural components in the cell (Finkel et al 2010). The second strategy is the ability to ‘make do with less’ which may be accomplished by physiological substitution of a P-containing lipid with a non-P-containing lipid (sulfolipid), and many cyanobacteria are able to do this (Van Mooy et al 2009). Thus the cellular C : P content of Synechococcus is about 100, whereas this ratio in a typical diatom is about 50 (Finkel et al 2010). The third strategy is the ability to acquire P in organic or particulate form, via alkaline phosphatase activity or mixotrophy, which may provide some cells a source of P not available to those cells dependent on inorganic P for their nutrition. Many dinoflagellates have a comparatively high cellular P requirement, and therefore the ability to consume particulate P may be an important reason why these types of cells can thrive when some others cannot. An added competitive benefit for these cells is that there may also be a growth advantage when feeding mixotrophically, compared to pure autotrophic growth (Jeong et al 2004, Glibert et al 2009, Flynn et al 2013), thus mixotrophy is a major mode of nutrition by HABs in eutrophic waters (Burkholder et al 2008) and may help to sustain blooms when dissolved nutrients are depleted.

Of particular concern is the association of increased toxin in many HABs under condition of elevated N : P availability. Many cyanobacteria and dinoflagellate toxins are N-rich compounds and thus these cells require a supply of N in order to synthesize these metabolites. The most ubiquitous cyanotoxin in freshwater systems are microcystins (MCs), hepatotoxic compounds that can be lethal to mammals if ingested (Carmichael 1994, Chorus and Bartram 1999), some congeners of which are significantly more toxic than others (Sivonen and Jones 1999). Microcystis (Chroococcales), Anabaena (Nostocales), and Planktothrix (Oscillatoriales) are among the taxa that can synthesize MCs (Cronberg and Annadottir 2006). Total concentrations of MC have been strongly related to N concentrations in several comparative studies (Giani et al 2005, Rolland et al 2005, Dolman et al 2012, Monchamp et al 2014). Although negative relationships between MC and N : P ratios have been reported for lake Tai (Taihu), China (Otten et al 2012), but in a cross system analysis of freshwaters (Orihel et al 2012), the authors of the findings of negative correlations emphasize that such results are relevant only to hypereutrophic conditions in which TN concentrations are also high (>100 μM N). Furthermore, a reanalysis of the aforementioned comparative study reported highest MC concentration at intermediate N : P ratios (Scott et al 2013). Excess N and high N : P ratios have also been related to MC production under controlled culture conditions (e.g., Lee et al 2000, Oh et al 2000, Vézie et al 2002, Downing et al 2005, Van de Waal et al 2009). Some in situ evidence, albeit weak, suggests that increased N availability may influence the MC congener type to more toxic variants that have higher N content (e.g., Van de Waal et al 2009) and that P limitation causes an increase in N-rich toxins of numerous HABs (Van de Waal et al 2014).

Some marine HABs also show increased toxin production under conditions of elevated N : P ratios. As examples, under conditions of elevated N : P, hemolytic activity per cell increases by up to an order of magnitude in the prymnesiophytes Prymensium parvum and Chrysochromulina (now Prymnesium) polylepis (Johansson and Granéli 1999), and neurotoxin production increases in the diatom Pseudo-nitzschia multiseries and in the dinoflagellates Karlodinium venificum, Alexandrium sp., and Karenia brevis (Granéli and Flynn 2006, Hardison et al 2013).

Recent reviews of the physiological bases of N uptake as well as molecular and metatranscriptomic data lend considerable support to the emerging conclusion that diatoms are specialists in use of oxidized forms of N, while cyanobacteria and dinoflagellates are specialists in reduced forms of N (e.g., Glibert et al 2014a and references therein). A considerable amount of experimental evidence supports the notion that freshwater cyanobacteria seem to favor reduced N forms (Blomqvist et al 1994, Berman and Chava 1999). Extensive laboratory molecular evidence of the use of both urea and NH₄⁺ by cyanobacteria (Flores and Herrero 2005), and gene expression data (Ginn et al 2009) also lend support to this conclusion. Mesocosm studies conducted in fresh and brackish systems further illustrate that when enriched with oxidized versus reduced forms of N, even when the TN supply remains the same, proportionately more diatoms are produced under oxidized N conditons while more cyanobacteria and cryptophytes are produced under conditions of increasing reduced forms of N (e.g., Finlay et al 2010, Donald et al 2013, Glibert et al 2014a). Donald et al (2011) found in mesocosm enrichment studies conducted in the Northern Prairie lakes of Canada, that urea additions stimulated M. aeruginosa growth but MC concentrations increased even more. Of the comparative field studies in freshwater, toxic cyanobacterial species appear to be favored over diatoms as a function of N availability in reduced relative to more oxidized forms in the hypereutrophic Taihu and Lake Okechobee (McCarthy et al 2009). In marine systems, cyanobacteria and dinoflagellates have also been associated with proportionately greater use of reduced forms of N and diatoms greater use of oxidized forms of N (e.g., Berg et al 2003, Heil et al 2007).

In situ evidence of the role of N forms in cyanobacterial community composition remain rare and influence on MC concentrations or congener composition even rarer. In Québec, Canada, the ratio of N : P in fertilizer acquisition increased steadily from 1977 to 1995, and accelerated from 1995 to peaks of >3 (on a weight basis) from 2006 to 2008 (figure 2(A)). There was a marked increase in the purchase of N fertilizer in the form of urea or products containing urea as of 2003 with a urea peak observed in 2007. That year was considered an ‘exceptionally favorable’ year for agricultural commodities (Heffer and Prud’homme 2008), and a sustained higher overall use of urea products has occurred since that time (figure 2(B)). Records of reported cyanobacterial incidences in lakes across the province (see SI) suggest that the peak in events occurred in 2007 (figure 2(C)), when both the N : P ratio and urea use were high. Since then the incidences of blooms have remained high (figure 2(C)). The correlation of incidences of blooms in these lakes with the increase in forms of N fertilizer containing urea (figure 2(D); r² = 0.42, p = 0.058) is suggestive of such an effect, although more years of data will be required to substantiate this relationship. While these patterns provide only indirect evidence, in a comparison over the course of the growing season of three lakes in Québec known to have toxic cyanobacteria, cyanobacterial community structure was primarily influenced by the availability of chemically reduced and organic N forms (DON and NH₄⁺) and temperature (Monchamp et al 2014). N forms and concentration however did not influence congener composition or the toxicity of the dominant variant in that study, but cyanobacterial community structure did. This suggests complex interaction between the availability of N and other environmental variables in influencing community structure with an indirect effect on congener composition and overall bloom toxicity.

Zoom In Zoom Out Reset image size

Several of the more potent marine HAB species show the same trend. For Alexandrium tamarense, the availability of urea has also been related to toxin content of the cells: the toxin content for urea-grown cells was found to be higher than that of NO₃⁻-grown cells, but not as high as cells grown on NH₄⁺ (Leong et al 2004). Furthermore the biosynthesis of toxin when grown on urea appears to differ from that which occurs under NO₃⁻ or NH₄⁺ growth conditions. For another dinoflagellate, Karenia brevis, up to six-fold increases in toxin content have been observed during growth with elevated urea availability compared to controls without urea enrichment (Shimizu et al 1993). For the toxic diatom, Pseudo-nitzschia sp., increases in toxicity in both laboratory cultures and natural field assemblages have also been found for cells growing on urea compared to those growing on NH₄⁺ or NO₃⁻ (Cochlan et al 2008, Kudela et al 2008). On a global scale there is a simultaneous increase of total fertilizer N, particularly in the form of urea, and the frequency and extent of a number of HAB cyanobacteria and dinoflagellate species providing further, though indirect, evidence for a relationship between urea and HABs (Glibert et al 2006, 2008).

China presents an interesting case study in terms of the relationships between increasing HAB frequency and the changes in fertilizer use and export. All data used herein come from pre-existing data sources, industry statistics and published literature (see SI). Fertilizer N use in China has escalated from about ∼0.5 MT in the early 1960s to 42 MT around 2010, with the fraction of urea increasing nearly five-fold over just the past two decades (figure 3(A); IFA 2014, FAO 2012, Zhang and Zhang, no date). River export of N increased from 1980 to 2010 from ∼500 to >1200 kg N km⁻² yr⁻¹ in the Changjiang River (Yantgtze River), from ∼100 to ∼200 kg N km⁻² yr⁻¹ in the Huanghe River (Yellow River), and from ∼400 to >1200 kg N km⁻² yr⁻¹ in the Zhujiang River (Pearl River) basins, the latter having one of the largest N : P ratios in the world (Ti and Yan 2013). The annual N load from the Changjiang River to the coastal ocean is higher than loads from the Mississippi and the Amazon rivers (Goolsby and Battaglin 2001, Duan et al 2008). Atmospheric sources of NH₄⁺, which is the dominant form of inorganic N on a regional basis, and which exceeds that of oxidized N forms in the Changjiang River basin, are due primarily to livestock excretion, fertilizer N use, and human waste (52%, 33% and 13% respectively, Xiao et al 2010).

Zoom In Zoom Out Reset image size

The number of HABs has increased in all waters of China in the past three decades and most inland and coastal waters are rated in the moderately to severely polluted range (Wang et al 2011, Ti and Yan 2013). In Taihu, the largest lake in the Changjiang watershed, the duration of cyanobacterial Microcystis blooms has increased from ∼1 month yr⁻¹ to nearly 10 months yr⁻¹ in the past 15 yr (Duan et al 2009; figure 3(B)). The Changjiang watershed produces >30% of China’s agricultural output (Xing and Zu 2002) and in the Taihu region, N fertilizers from manures and mineral fertilizers have increased eight- to ten-fold since the 1950s (Chen et al 2008) and fertilizer N use in Jiangsu Province (2.5 MT N yr⁻¹) is 80% of TN inputs (PBL 2012). The change in HABs in Taihu is strongly related to the increase in urea and in the ratio of urea : P₂O₅ use scaled to the Changjiang watershed (Chen et al 2003, Ye et al 2007, Duan et al 2009; figures 3(B)–(D); r² = 0.85 and 0.92, respectively, p <0.01). These trends in availability of reduced forms of N and in increased N : P ratios in this region are further accelerated by aquaculture development and atmospheric deposition. In 2007, 13% of Chinese inland aquaculture production was in Jiangsu (China Fishery Yearbook 2007), and large quantities of NH₄⁺ and urea with high N : P ratio are also discharged from that source to Taihu. It should be noted that waste from aquaculture not only has a high N : P ratio (>20) but is largely in reduced N form with significant amounts of urea (Bouwman et al 2013). Also, 80% of the atmospheric deposition in this region has been associated with NH₄⁺ mobilization from intense fertilization during the growing period (Chen et al 2008). With intensive agriculture, N deposition onto the lake is a direct N input not accompanied by P.

In Chinese marginal seas there are similar trends of increased HABs over the past several decades (Wang et al 2008, 2011; figure 4(A)). When all China’s marginal sea bloom occurrences are considered, there are strong relationships between both number of HAB reports and aerial HAB extent and N use as urea (figures 4(B) and (C); r² = 0.53 and 0.37, respectively, p < 0.01).

Zoom In Zoom Out Reset image size

In both the Huanghai (Yellow) Sea region and in the Northern South China Sea, the number of reported occurrences of HABs has increased in parallel with the long-term trend in N : P of the water column (figures 4(D) and (E); r² = 0.46 and 0.29, respectively, p = 0.06). As in the Taihu example, these relationships do not account for nutrient input from coastal aquaculture or from atmospheric NH₄⁺ deposition that would enhance these relationships. Coastal N release from Chinese mariculture increased from 2000 to 2010 by 45% on average and up to 63% in some provinces while P increase was only 37% (Bouwman et al 2013). In 2004, NH₄⁺ and organic N comprised 75% of N in wet atmospheric deposition in Southeast China (Chen et al 2011).

At the mouth of the Changjiang River, molar NO₃⁻ : PO₄³⁻ values were ∼30–40 in the 1960s but had risen to >250 by the late 1990s (Shen and Liu 2009; figure 4(D)). Occurrences of HABs in the East China Sea were rare in the 1970s, had increased to a few dozen over the 1980s but had increased to >130 in just the 5 yr from 1992 to 1997 (Shen and Liu 2009) but the scale of the blooms soared in the years since; the spatial scale of the annual blooms increased from 1000s of km² in 2000 to >15 000 km² by 2005 with many millions of dollars lost in high value aquaculture products due to associated fish kills (Li et al 2009). The dinoflagellates Prorocentrum donghaiense and Karenia mikimotoi are among the common HABs now reported in East China Sea (e.g., Zhou et al 2008, Li et al 2009).

In the Huanghai Sea region, inorganic N : P ratios are now about twice Redfield proportions, and about four-fold higher than in the 1990s (figure 4(E)). There has also been nearly a six-fold increase in HAB occurrences and a shift to proportionately more dinoflagellates compared to diatoms (Fu et al 2012). In the South China Sea region, water column inorganic N : P ratios increased from ∼2 in the mid-1980s to >20 in the early 2000s (Ning et al 2009). In addition to the increase in number of HABs, a change in species composition to increasing dominance of species such as Chattonella, Gymnodinium breve, and Dinophysis has occurred (Wang et al 2008).

The results herein lend support the view that both N and P controls are necessary to reduce eutrophication in both freshwater and marine waters (e.g., Burkholder et al 2006, Howarth and Paerl 2008, Conley et al 2009, Paerl 2009, Glibert et al 2011, 2013). Loss of biodiversity, and effects on ecosystem and human health due to eutrophication are considered major challenges of our current day (e.g., Borja 2014). Given the known inefficiencies in all aspects of N use in both industrialized nations and throughout the world, the benefits of increased emphasis on N reduction, and improved N use efficiencies at all levels of the production side, would have far reaching benefits to ecosystems and especially to water quality (Houlton et al 2013, Sobota et al 2013). Calculated damage costs of loss of ecosystem services, eutrophication and human health due to loss of reactive N to the environment are large (Compton et al 2011, Sobota et al this issue). So too are the economic impacts to aquaculture, increased risks to human health and ecosystems, and losses to fisheries and ecosystem services due to HABs (Hoagland and Scatasta 2006). More work is needed to accurately quantify the sources and fluxes of N at all stages of the supply-to-loss pathways, and there is no question that control of N may be more challenging than control of P. Given our increasing knowledge of the physiological response of many HABs to increasing N : P and to increasing ratios of reduced : oxidized forms of N, we can conclude that without more aggressive N control the future outlook in terms of HABs is likely to include more events, more often, and such events may also be more toxic (O’Neil et al 2012, Glibert et al 2013, 2014b, Paerl and Otten 2013). Indeed, a recent modeling effort in which the physiological responses of HABs to altered N : P ratios and altered N form, even without any further increases in N loading, in conjunction with projected climate change effects, suggest an expansion in area and/or number of months annually conducive to development of several HAB genera along the NW European Shelf-Baltic Sea system and NE Asia by end of the century (Glibert et al 2014b). Such projections alone should be cause for advancing our understanding of the relationships between HABs and nutrient loading, and together with the projected global expansion in N loading should be serve to sound the alarm that our existing approaches to nutrient management of sensitive coastal and freshwaters are not sufficient, particularly in the face of climate change and other stressors.

Acknowledgements