Joseph J. Donovan

DROUGHT CYCLES AND LANDSCAPE RESPONSES TO PAST ARIDITY ON PRAIRIES OF THE NORTHERN GREAT PLAINS, USA

Widespread drought is among the most likely and devastating consequences of future global change. Assessment of drought impacts forecast by atmospheric models requires an understanding of natural drought variability, especially under conditions more arid than today. Using high-resolution lake-sediment records from the northern Great Plains, we show pronounced 100- to 130-yr drought cycles during the arid middle Holocene (8000 calendar yr BP). During drought phases, grass productivity declined, erosion and forbs increased, and fuel limitation reduced fire importance. Intervening humid decades saw grass production rise, with stabilization of soils and renewed fire as fuels became abundant. Although both C3 and C4 grasses declined during droughts, a lasting shift to C3 dominance occurred during a single drought -8200 calendar yr BP. During the more humid Late Holocene (2800 calendar yr BP), climate was less variable and without evident drought cyclicity. Consequently, drought severity during past, and possibly future, arid phases cannot be anticipated from the attenuated climate variability evident during contemporary humid phases. Our study demonstrates that agriculturally important grassland ecosystems respond sensitively to drought variability, uncertainty in which has profound implications for the future of these ecosystems.

Ground-water processes controlling a prairie lake's response to middle Holocene drought

Lake response to climate forcing is commonly regarded as being controlled by within-lake processes. We present a case from the northern Great Plains in which Holocene drought response recorded in lake sediments was determined by changes in chemical characteristics of source ground water. Middle Holocene ostracode assemblages were dominated by halophylic Limnocythere staplini , indicating increased lake-water salinity, yet the oxygen isotope values in this period decreased by 2‰ to 3‰, contrary to expectations of evaporative dominance. We infer from these proxy records and ground-water flow-model results that the paleoclimatic record in the lake was mediated by changes in ground-water catchment area and by reduction in evapotranspiration accompanying desiccation of peripheral lake basins.

Holocene dynamics of the Florida Everglades with respect to climate, dustfall, and tropical storms

Significance Wind-blown dust is seldom considered an important source for nutrients in large peatlands, such as the Everglades. However, a sedimentary record suggests that high loadings of dust-borne nutrients once prevailed in the central Everglades during a period of moister climate with intense tropical storms that ended 2,800 y ago. Afterwards, a drier climatic regime with a steep decline in dustfall may have been the impetus for the striking surface patterning of the Everglades. This study provides additional support for the importance of aeolian dust in ecosystem development. Aeolian dust is rarely considered an important source for nutrients in large peatlands, which generally develop in moist regions far from the major centers of dust production. As a result, past studies assumed that the Everglades provides a classic example of an originally oligotrophic, P-limited wetland that was subsequently degraded by anthropogenic activities. However, a multiproxy sedimentary record indicates that changes in atmospheric circulation patterns produced an abrupt shift in the hydrology and dust deposition in the Everglades over the past 4,600 y. A wet climatic period with high loadings of aeolian dust prevailed before 2800 cal BP (calibrated years before present) when vegetation typical of a deep slough dominated the principal drainage outlet of the Everglades. This dust was apparently transported from distant source areas, such as the Sahara Desert, by tropical storms according to its elemental chemistry and mineralogy. A drier climatic regime with a steep decline in dustfall persisted after 2800 cal BP maintaining sawgrass vegetation at the coring site as tree islands developed nearby (and pine forests covered adjacent uplands). The marked decline in dustfall was related to corresponding declines in sedimentary phosphorus, organic nitrogen, and organic carbon, suggesting that a close relationship existed between dustfall, primary production, and possibly, vegetation patterning before the 20th century. The climatic change after 2800 cal BP was probably produced by a shift in the Bermuda High to the southeast, shunting tropical storms to the south of Florida into the Gulf of Mexico.

Lateglacial and Holocene hydroclimate inferred from a groundwater flow-through lake, Northern Rocky Mountains, USA

Climate-driven variations in lake-groundwater exchange are recorded by sediments in groundwater-dominated lakes. A groundwater flow-through lake in west-central Montana (USA) registers latest Pleistocene and Holocene hydroclimatic variation in fluid and solute balance, as controlled by rates and timing of groundwater recharge. Early Holocene warming occurred under conditions of relative aridity and low groundwater throughflow, punctuated by a c. 450-yr episode of lake dilution centered on 11 000 cal. yr BP. Maximum evaporative concentration of lake waters, registered in both δ18O values and mineralogy of endogenic carbonates, coincided with the early-Holocene peak in insolation seasonality at about 9750 cal. yr BP. Subsequently, progressively decreasing lake residence time drove a sustained long-term decline in salinity while having a very subdued effect on mean δ18O values. We explain this decoupling by (1) limits placed on oxygen isotope sensitivity by groundwater throughflow, and (2) a shift toward greater summer rain contribution to lake inflow after mid-Holocene time. Superimposed multidecadal- and century-scale variation in lake—groundwater exchange generated high-frequency but low-amplitude isotopic oscillations throughout the record. High rates of groundwater throughflow maintaining low lake salinity similar to that observed today were established around 1400 cal. yr BP. We infer reduced regional stream baseflow, decrease in permanent wetlands (relative increase in ephemeral wetlands) and enhanced lake and wetland salinity prior to this time, relative to the late Holocene.

Environmental history of a closed-basin lake in the US Great Plains: Diatom response to variations in groundwater flow regimes over the last 8500 cal. yr BP

Sediment records from closed-basin lakes in the Northern Great Plains (NGP) of North America have contributed significantly to our understanding of regional paleoclimatology. A high-resolution (near decadal) fossil diatom record from Kettle Lake, ND, USA that spans the last 8500 cal. yr BP is interpreted in concert with percent abundance of aragonite in the sediment as an independent proxy of groundwater flow to the lake (and thus lake water level). Kettle Lake has been relatively fresh for the majority of the Holocene, likely because of the coarse substrata and a strong connection to the underlying aquifer. Interpretation of diatom assemblages in four groups indicative of low to high groundwater flow, based on the percent of aragonite in sediments, allow interpretations of arid periods (and probable meromictic lake conditions) that could not be detected based on diatom-based salinity reconstructions alone. At the centennial–millennial scale, the diatom record suggests humid/wet periods from 8351 to 8088, 4364 to 1406 and 872 to 620 cal. yr BP, with more arid periods intervening. During the last ~ 4500 years, decadal–centennial scale periods of drought have taken place, despite the generally wetter climate. These droughts appear to have had similar impacts on the Kettle Lake hydrology as the ‘Dust Bowl’ era droughts, but were longer in duration.

Episodic struvite deposits in a Northern Great Plains flyway lake: indicators of mid-Holocene drought?

A series of 34 layers of struvite (hydrous Mg-ammonium-phosphate) were deposited in laminated sediments of Kettle Lake, North Dakota, in the Northern Great Plains, USA. Sedimentologic, mineralogic and nitrogen isotopic evidence suggest that struvite was deposited during nutrient-enriched eutrophication events within a Mg-rich lacustrine environment. The struvite layers are dispersed between 4650 and 8700 cal. yr BP, with particularly high frequency between 8068 and 8700 cal. yr BP. The youngest struvite layer dates to 2734 cal. yr BP. Key features of the struvite-forming events were (a) relatively low water stage and consequent elevated water salinity associated with the dry mid Holocene, (b) disappearance of most lakes and wetlands in the region, and (c) focused, but brief, visitations by large populations of migratory waterfowl. The lack of more frequent occurrence of struvite is ascribed to the rarity with which this combination of conditions was achieved, in combination with difficulty of preservation for this salt. 15N isotopes in struvite (mean 7.51‰) are heavy relative to 15N in sediment and also extremely uniform over time. The isotopes are interpreted to reflect a waterfowl waste source without extensive NH4 volatilizaton. The timing of the struvite events does not closely correspond to century-scale mid-Holocene drought cycles inferred from oscillating aragonite concentrations, although there is a weak preferred occurrence in the humid phase of these cycles. Thus the struvite events are ephemeral prairie features of this generally arid period but tend not to occur at precisely the most arid intervals.

Geochemical evolution of lacustrine brines from variable-scale groundwater circulation

Evaporative groundwater-fed lakes in the glaciated North American Great Plains vary widely in chemistry. A contributing cause is chemical variability of source groundwater intercepted by specific lakes, caused in part by differing depths of groundwater circulation. Aqueous chemical characteristics of 61 lakes and 160 groundwater samples were compared for an area where such lakes are common in eastern Montana-western North Dakota. Results indicate that groundwater chemistry varies according to depth in a similar fashion within different aquifers. Lake water evaporation from initial groundwater solutions typical of three depths was geochemically modeled using PHRQPITZ, based on a Pitzer treatment of activities and equilibria. Results show that chemistry of most lake waters in the study area may correspond to that predicted from evaporation of shallow- and intermediate-depth groundwater, but not of deep groundwater as postulated in some previous investigations. Lakes in shallow surface depressions receive water primarily from shallow (local) groundwater flow; lakes located in deep or broad topographic depressions may additionally receive groundwater from deeper circulation. In the field area studied, relative dominance of anions (sulfate vs. carbonate) in brines is a signature for inferred depth of source. Also diagnostic is the suite of brine salts formed (NaSO4Mg salts for shallow flow; these plus NaCO3 salts for intermediate depth flow). Such source signatures will vary from area to area according to depth variations in groundwater chemistry and in stratigraphy. Chemical evolution of lake water is a two-stage process, with a groundwater path (influenced by residence time, depth of circulation, aquifer mineralogy, and related factors) and a surface path (influenced by evaporation rates, lake-aquifer hydraulics, and lake geochemical reactions). Groundwater flow patterns may affect the former set of factors, thereby indirectly controlling lake water chemistry.

Double-flow behavior observed in well tests of an extremely heterogeneous mine-spoil aquifer

Abstract A radial-flow investigation was performed to examine hydraulic characteristics of a heterogeneous unconfined mine-spoil aquifer formed by reclamation of surface-mine overburden. Slug test estimates of hydraulic conductivity (K) range over four orders of magnitude (from 10−6 to 10−2 m/s) and display two prominent statistical modes, ascribed to fine matrix and coarse rubble zones, respectively. A large proportion of wells displayed either of two anomalous slug-test response patterns, attributable to zones of highK around or near the well. Slug-test estimates of matrixK appear unbiased, but those of high-conductivity zones are thought to contain error induced by local heterogeneity and drainage of storage from near the water table. Numerical simulation of slug tests suggests that high-K zones, vertically continuous to the water table and either in the annulus of (e.g. well skin) or in close proximity to the well, may produce response similar to that observed. It is not possible to distinguish well-skin from heterogeneity effects based on the nature of slug test response alone. Pumping yields, however, are generally higher for wells with anomalously rapid slug-test response and suggest that, in some cases, the heterogeneity extends to large distance from the well. Despite their local region of influence, slug tests in such extremely heterogeneous aquifers may encounter highly conductive zones that may reduce the reliability of their results if they permit water-table drainage to the well.

Geochemical response of acid groundwater to neutralization by alkaline recharge

Solute transport and chemical neutralization (pH 3 to 7) within a shallow heterogeneous aquifer producing acid mine drainage (AMD) are examined at an abandoned surface coal mine in West Virginia. The aquifer is undergoing partial neutralization by mixing with alkalinity from a leaking sludge disposal pond, extending in preferential zones controlled by aquifer heterogeneity. Hydraulic heads interpolated from wells indicate leakage from a central alkaline (pH 7.1, 0.72 meq/L alkalinity) sludge pond is a principal source of recharge. Chemically-conservative sodium, added to AMD during treatment and leaked into the aquifer with the sludge, develops a dispersion plume over a restricted portion of the aquifer that correlates with pH, hydraulic head, and dissolved metals distributions. Concentrations of aluminum, iron, sulfate and acidity display higher concentrations downgradient from the pond as sludge alkalinity is consumed along flow paths. Before reaching springs, most dissolved iron is oxidized and hydrolyzed, likely precipitating in the aquifer as a ferric hydroxide or hydroxysulfate phase. The spatial pattern of iron and aluminum concentrations suggests accelerated oxidation caused by gas transport along the outer slopes of the spoil. Dissolved aluminum concentrations increase with total acidity, suggesting that dissolution of silicate minerals results from acidity released by iron hydrolysis. Neutralization reactions and higher pH are favored in more highly permeable portions of the spoil, where ferrihydrite and aluminum hydroxysulfate minerals (such as basaluminite) are supersaturated. In acid-producing zones at pH < 4.5, jurbanite is near equilibrium and an aluminum-sulfate phase with similar properties may limit aluminum concentrations, but become undersaturated in zones of advancing neutralization. At this particular site, ferrous iron produced by pyrite oxidation is almost completely oxidized over short transport distances, allowing hydrolysis of iron and aluminum should sufficient alkalinity be added to these acid waters.

Selenium Adsorption onto Iron Oxide Layers beneath Coal-Mine Overburden Spoil.

A field experimental study to determine the feasibility of sequestering dissolved selenium (Se) leached from coal-mine waste rock used an iron (Fe)-oxide amendment obtained from a mine-drainage treatment wetland. Thirty lysimeters (4.9 × 7.3 m), each containing 57.7 t (1.2-1.8 m thickness) of mine-run carbonaceous shale overburden, were installed at the Hobet mine in southeastern West Virginia. The fine-grained Fe-oxide was determined to be primarily metal oxides (91.5% ferric and 4.37% aluminous), with minor (<3%) SO and Ca, perhaps as gypsum. The mineralogy of the Fe was goethite, although residual ferrihydrite may have been present. Various thicknesses of this amendment (0.0064, 0.057, 0.229, and 0.457 m, plus a zero-amendment control) were used, ranging from 0 to 2.2% weight percent of the spoil. The control and each treatment were replicated six times to estimate uncertainty due to compositional and hydrological variation. Infiltration of rainfall created leachate that drained to individual batch-collection tanks that were sampled 46 times at approximately 2-wk intervals from 2010 to 2012. Basal Fe-oxide layers in the three highest amendment categories removed up to 76.1% selenium (in comparison to unamended piles) from leachate by adsorption. Only lysimeters with very thin Fe-oxide layers showed no significant reduction compared with unamended piles. Reproducibility of replicates was within acceptable limits for amended and unamended lysimeters. Results indicate that in situ amendment using Fe-oxide obtained from treatment of mine water can sequester Se by adsorption on surfaces of goethite and possibly also ferrihydrite. This process is demonstrated to substantially reduce dissolved Se in leachate and improve compliance with regulatory discharge limits.