Flood Damage on Agricultural Land and Methods for Restoration of Agricultural Soils after Catastrophic Floods in Cold Areas

Abstract

Catastrophic floods have large effect on agricultural land both in short and long term. In this chapter, examples of impact of floods of different size in cold regions with glaziers have been presented. The largest floods occur as combination of heavy rainfall and melting and snow and ice in the mountainous areas. Periods of waterlogging by cold running water resulted in decreased yields, but N-fertilization after the soil no longer was water saturated could reduce the yield loss considerably. Although the floods cause severe erosion and sedimentation, results show that it is possible to find measures for reconstruction of the soils with the same productivity as undamaged soils, while the average result was about 85% of the original productivity.

1. Introduction

In recent years, large and catastrophic floods have received much attention. It has been claimed that the large floods occur more often than previously, and the large floods have a more catastrophic pattern. A wilder and wetter climate due to climatic change has been predicted, causing more severe floods with large damage on infrastructure, agricultural soils, loss of lives of animals and humans. In Europe, the number of catastrophic floods was reduced in the period 1870–2016, while the number of humans losing their lives during catastrophic floods increased [1]. A timespan of about 150 years of measurements is too little in order to verify the effects of the largest floods, which may occur once in a millennium or less often. The mechanisms behind the largest floods may be of interest to understand in order to make models for future climatic change.

In many countries, fluvial plains are important agricultural areas with high productivity, but these areas are also the areas that are most exposed to floods. The pattern of Fluvisols [2] with sediments that receive fresh material or have received it in the past and still show stratification indicates areas exposed to floods. Large floods make severe damage on agricultural land by directs damage on crops, erosion, and sedimentation, and different methods for evaluation of flood damage to agriculture have been developed [3]. However, measures for land reclamation and restoration of damaged agricultural areas have sparsely been reported in scientific literature and seem to be absent for areas where combination of snowmelt, melting of glaziers, and heavy precipitation occur.

The aim of this chapter is to show the importance of large floods and the impact on agricultural land both in a historical context and more recent examples with a special focus on restoration and reclamation measures after catastrophic floods in cold regions. The study area is the two largest valleys in Norway; Gudbrandsdalen with the river Gudbrandsdalslågen and Østerdalen with the river Glomma in central eastern Norway, which was severely damaged by a flood in June 1995 [4]. The same area was hit by the largest flood in historical time in July 1789 [5], while the geomorphology of the valleys has been formed by large outburst floods from glacial lakes about 10,000 years before present [6].

2. Material and methods

After the 1995 flood in central eastern Norway, research projects were started in order to document the direct damage of the flood by field experiments at flooded areas. After the direct effects were studied, the effects of different restoration measures were investigated. One of the areas with largest damage after the 1995 flood in the river Glomma was Øksna, north of Elverum in Østerdalen (60°58′ 15″ N, 11°29′ 05″ E). The restoration work took place in the period 1996–1998, and an evaluation of the different measures was done based on studies in a 4 year period, 1999–2002. In total, 11 different measures, including undamaged references, were studied at 26 plots (Table 1, Figure 1).

Type No.	Damage	Restoration	Texture	Plot No.
1	Erosion crater, >2 m deep	Filled with sand	Loamy fine sand	1, 4, 7
2	Eroded topsoil	Leveling of topsoil	Sandy loam	2, 5, 8
3	Slightly eroded topsoil	Normal tillage	Sandy loam	3, 6, 9
4	Erosion crater,; > 2 m deep	Filled with sand, limed sewage sludge mixed in topsoil	Loamy fine sand	10, 11, 12
5	Intact topsoil	Normal tillage	Loamy sand	13, 26
6	Small damage	Leveled, topsoil replaced	Sandy loam	14, 25
7	Erosion crater	Filled with till and sand in bottom, sandy loam as topsoil, sewage sludge mixed in topsoil	Sandy loam	15, 24
8	Small damage	Leveled, topsoil replaced	Sandy loam	16, 23
9	Erosion crater	Filled with sand, sewage sludge mixed in topsoil	Loamy fine sand	17, 22
10	Erosion crater	Filled with till and sand in bottom, sandy loam as topsoil, sewage sludge mixed in topsoil	Sandy loam	18, 21
11	Erosion crater	Filled with till and sand in bottom, sandy loam as topsoil, peat mixed in topsoil (upper 10 cm)	Loamy fine sand	19, 20

Table 1.

Flood damage and restoration measures at Øksna, Elverum commune, Norway.

Figure 1.

Plots for soil and yield investigations at Øksna, Elverum commune, Norway. Photo from 1998 [7].

Two different types of sewage sludge were used as soil conditioner, a limed sewage sludge with pH >10 (Plots 10–12), and a sewage sludge where aluminum salts had been used for coagulation for removal of P from wastewater. The amount of sludge used was about 20 Mg dry matter ha⁻¹, which is the permitted amount according to Norwegian regulation on organic fertilizer.

The yield of cereals, spring barley (Hordeum vulgare) and oats (Avena sativa), was monitored at the plots in the period 1999–2002. Analysis of variance (ANOVA) was performed in order to document if there were significant crop yield differences between the three groups of treatments: eroded topsoil or intact topsoil, leveled areas, and filled erosion craters. As there were unequal sizes of the groups, the Bonferroni method for multiple comparisons was used (p < 0.05).

Soil properties were investigated at plot level, including texture, pH (H₂0), total organic C, representing the topsoil (Ap-horizon). Readily available plant nutrients were determined according to the AL-method [8]. Intact soil cores (100 cm³) were taken for determination of bulk density and total porosity at nine plots with three cores per plot in the topsoil (5–9 cm depth). Total porosity was measured as water content at saturation, while bulk density was determined after drying the soil cores 24 h at 105^°C.

3. Effects on flooding on crops and restoration of flood damage

In the period from 22th May to 2nd June 1995, large part of central eastern Norway was hit by a severe flood, in media called a catastrophic flood. In the period from 22th to 25th May, the temperature rose by 5–10°C causing enormous snowmelt in the mountainous region. It was estimated that 4000 million m³ snow melted during the period from 25th to 2nd June, which is equivalent to 100 mm precipitation distributed over the whole watershed. In addition, 50–70 mm precipitation as rain hit the central part of the watershed during the days from 28th May to 1st June. This caused a large flood in the rivers Glomma in Østerdalen and the river Gudbrandsdalslågen in Gudbrandsdalen. The maximum observed discharge in Glomma was close to 3100 m³ s⁻¹ at Elverum [4].

During the 1995 flood, about 14,000 hectares agricultural land was flooded, of which about 10,000 hectares along Glomma river in Østerdalen. It has been estimated that 2 million m³ was eroded by the flood, while the amount of sediments overlying soils constituted of at least 1 million m³. About 1000 hectares was seriously damaged either by erosion or sedimentation and could not be used without restoration measures [4].

3.1 Crop damage of flooding and effects of fertilization

As the flood in central eastern Norway happened after sowing in the start of the growing season in 1995, it was possible to make research on crop damage and measure if fertilization after the flooding could make reasonable crop yields.

It was found that potatoes had almost total crop failure after waterlogging, even if the period of waterlogging lasted only 1–2 days. Areas with grass for silage and hay production had delayed growth, but recovered after waterlogging and gave reasonable yields [9]. Waterlogging of fields with cereals gave significant yield loss, but the duration of the waterlogged period had less impact (Table 2).

Waterlogged	Solør			Selsvollene
	Yield kg ha^−1	Rel. yield	Protein %	Yield kg ha^−1	Rel. yield	Protein %
0 days	620	100	14.1	762	100	13.9
1–2 days	398	64	13.9	627	82	13.7
4–5 days	371	60	13.2

Table 2.

Influence on waterlogging on yield (kg ha⁻¹, 15% water content) and protein (% of DM) of spring wheat (Triticum aestivum) on fields in Østerdalen (Solør) and Gudbrandsdalen (Selsvollene) in 1995.

As the flooding caused leaching of soluble plant nutrients from the soil and fertilizer, which had been applied in the spring, a series of field experiments were placed at waterlogged locations in Gudbrandsdalen and Østerdalen in order to measure the effect of fertilization after the soil was no longer water saturated [9].

Fertilization with N fertilizer after the fields had recovered sufficient bearing capacity after waterlogging had very good effect on crop yield (Tables 3 and 4). Although the total yield could be brought up to the level of not waterlogged soil by the extra dose of N-fertilizer >105 kg N ha⁻¹ (Table 4), the flooded areas had later maturing and up to three generations of tillers. There was no effect of other plant nutrients than N (Table 3).

Fertilizer	Yield kg ha^−1	Yield relative	Protein, %	Kg N in cereal yield, kg N ha^−1
None	3290	100	10.3	46
30 kg N ha^−1 CaNO3	3970	121	10.5	57
60 kg N ha^−1 CaNO3	4690	143	11.4	73
60 kg N ha^−1 CaNO3; + 28 kg K ha^−1 KCl	4650	141	11.2	71
60 kg N ha^−1 Yara Fullgjødsel® 21–4-10	4640	141	10.9	69
90 kg N ha^−1 CaNO3	5080	154	11.9	82
LSD, 5%	250		0.3

Table 3.

Effects of fertilization on yield of cereals, spring barley and oats (15% water content), protein content (% of DM), and uptake of N after 2–9 days of waterlogging during the flood in 1995. Means of 14 fields in Østerdalen and Gudbrandsdalen, Norway, modified after [9].

Treatment	Extra N by farmer, kg ha^−1	Extra N, experiment, kg ha^−1	Grain yield, barley, kg ha^−1
Not waterlogged	0	0	6750
Waterlogged 2–3 days	0	0	3960
Waterlogged 2–3 days	0	30	4610
Waterlogged 2–3 days	0	60	5940
Waterlogged 2–3 days	0	90	6430
Waterlogged 2–3 days	45	0	5270
Waterlogged 2–3 days	45	30	6220
Waterlogged 2–3 days	45	60	6960
Waterlogged 2–3 days	45	90	7180

Table 4.

Effects of N-fertilization after flooding on yield of barley (kg ha⁻¹, 15% water content) at Selsvollene in Gudbrandsdalen, Norway.

3.2 Restoration of areas damaged by erosion

The cereal yields varied much between the different types of restoration measures at Øksna. When the measures were grouped into three: eroded topsoil or no damage (Types No. 2, 3, 5), leveled areas (Types No. 6 and 8), filled erosion craters (Types No. 1, 4, 7, 9, 10, 11), significant difference was found between the undamaged areas (eroded topsoil/no damage) and the restored areas. Average of the restoration measures gave 85% of the yields compared to the areas with eroded topsoil or intact soil (Table 5). There was large variation within the group filled erosion craters. Type No. 4 gave 5340 kg ha⁻¹ yield of cereals as mean for the period 1999–2002 and type No. 11 gave only 2920 kg ha⁻¹ cereal yield in the same period. The results indicated that it is possible to find combination of treatments that can fully restore the productivity after severe erosion caused by flooding.

Treatment	Type No.	Yield of cereals, kg ha^−1
Eroded topsoil or intact topsoil	2, 3, 5	5280a
Leveled areas	6, 8	4450b
Filled erosion craters	1, 4, 7, 9, 10 11	4520b

Table 5.

Average yield of cereals (spring barley and oats) (kg ha⁻¹, 15% water content) at plots with different restoration measures at Øksna, Elverum, for the period 1999–2002. Means followed by different letters are statistically significant (p < 0.05).

The two types of sewage sludge had different effects on the yields. The limed and calcareous type had positive effect. The types Nos. 1 and 4 were filled with the similar loamy fine sand [Table 5], while sewage sludge was mixed in the topsoil at No 4. Although the difference was not statistically significant, the average yield for No. 4 was about 500 kg ha⁻¹ higher in the period 1999–2002 than for No. 1. Most of the treatments had low levels of readily available P in the topsoil (<5 mg 100 g⁻¹), which may result in P-deficiency for the crop [10]. The limed sewage sludge increased the level of readily available P in the topsoil significantly, resulting in very high levels of P-AL (>35 mg 100 g⁻¹). The other type of sewage sludge gave no additional yield compared to the other types of restoration measures. Although both sludges contained much phosphorus, soil analyses indicated deficiency of phosphorus after application of the sludge with Al salts used as coagulants. The plant availability of P in sludges produced from wastewater using Al salts as coagulants is often low compared with that in mineral fertilizer, while liming can increase the P fertilization effect of sludges [11]. This may explain the difference in effect between the two types of sewage sludge applied at Øksna.

There was no significant difference in TOC in the topsoil between areas with intact or eroded A-horizon and areas where the topsoil was amended with sludge during the restoration and the content of TOC was rather low, mainly 1 ± 0.3 g 100 g⁻¹ (Table 6). The type No. 11 where peat material had been mixed in the topsoil had low pH initially and was limed before the growing season 2000. Although this treatment had good physical properties for plant growth, this treatment had the lowest content of available plant nutrients and different plant deficiency symptoms were observed on the cereal plants (P, K, Zn) resulting in low yield of cereals.

Type No.	Sand, 0.06–2 mm	Silt, 0.002–0.06 mm	Clay, <0.002 mm	Bulk density, Mg m^−3	Porosity, m^3 m^−3	TOC, g 100 g^−1	pH (H20), 1999	pH (H2O), 2002
1	58.5	36.8	4.8	1.51	0.41	0.87	6.3	6.0
2	41.4	52.0	6.5	1.32	0.47	0.97	6.3	6.3
3	39.5	54.0	6.5	1.48	0.41	1.20	6.8	6.3
4	53.4	41.3	5.2	n.d.	n.d.	1.27	7.0	7.8
5	61.7	29.3	9.0	1.32	0.49	1.25	6.4	6.4
6	21.0	68.5	10.6	1.15	0.53	1.15	6.1	6.5
7	21.2	72.7	6.1	1.40	0.48	0.85	6.4	6.8
8	36.5	56.4	7.0	n.d.	n.d.	1.00	6.5	6.5
9	57.4	38.2	4.3	1.32	0.47	0.45	5.7	6.4
10	43.9	50.5	5.6	1.39	0.46	1.75	6.2	6.1
11	70.3	26.4	3.4	0.65	0.68	2.00	5.0	7.0

Table 6.

Soil texture, bulk density, porosity, TOC, and pH in topsoil (Ap-horizon) at Øksna for different restoration types.

4. Discussion

The tolerance to waterlogging varies between different cereal crops. Cereals as barley and wheat may produce adventitious roots with about 20% of aerenchyma [12]. Photosynthesis may continue under waterlogged conditions. In the present study, waterlogging up to 5 days reduced the relative wheat yield to 60–82% of the not waterlogged reference. This is within the range of yield decrease reported in literature [13]. Flooding by cold running water may cause the soil not to be anoxic, and the yield loss could be decreased compared to situations with higher water temperature and stagnant water [14]. During the large flood in 1995, cold melt water combined with rain gave relatively low temperatures in the water. Therefore the crops could survive for some days under waterlogged conditions. In 1995, the mean day and night temperature was below 6°C until 21st May and during the flood, the maximum day and night temperature was 15–18°C in the lowland parts and 8–10°C in the mountain area [4].

As the cereal yields could be brought up to the level of areas without waterlogging by sufficient N-supply after the flood, it seems reasonable that the period of flooding had delayed the growth due to the cold water. Nitrogen had most likely been lost by leaching during the flood, while denitrification may have occurred at sites with stagnant water.

The flood in July 1789 occurred as a result of 3 days of continuous rain combined with snowmelt of snow and ice in the mountainous areas. This led to a combination of landslides from the hillslopes and enormous flood discharge in the large rivers Gudbrandsdalslågen and Glomma and also in smaller side rivers [15]. Prior to this catastrophic flood, there had not been any large snow melt flood for many years and very cold winters. This can be related to the eruption of the volcano Laki at Iceland in 1783–1784. The effects of the aerosols and extreme volcanic pollution (i.e., dry fog) that effected Europe and other regions in 1783 have been estimated to cause a drop in temperature of 1.3°C in Europe and North America, lasting for 2–3 years [16]. After the flood in July 1789, large amounts of sand (50–100 cm thick) had sedimented above fertile soil with sandy loam texture in several places in Gudbrandsdalen. These areas gave very low yields and were subjected to drought. A deep plowing project showed promising results when the old plow layer was brought back to the soil surface. However, deep plowing gave only stripes with good effect [17], while use of excavator for bringing the old fertile soil to the surface fully restored the productivity of the soil. ¹⁴C dating of buried peat material below the sediments from the 1789 flood gave interesting results. In the peat layers it was found layers and lamina of silt up to 6–7 cm thickness between the organic layers. The difference in age between the bottom peat layer and the layer just below the thick mineral sediments was 4500–5000 years [5]. This indicates that the 1789 flood was an exceptional event, which had not happened for several millennia.

The reconstructed soils at Øksna and other areas damaged by erosion and sedimentation by the 1995 flood have been protected by building higher levees and similar measures, which prevented damage by a smaller flood in 2013. When agricultural lands close to large rivers are subjected to flooding in spring almost every year, the flood situation will have large impact on the ground water table in the soils at fluvial plain. Periods with high groundwater level may have large impact on the yields both on cultivated Fluvisols and reconstructed soils along the rivers, as found in a study in Nedre Eiker along the River Drammenselva, north of the city of Drammen, Norway [18].

Outburst floods from glacial lakes about 10,000 years BP have significantly influenced the geomorphology of the main valleys in eastern Norway, Østerdalen and Gudbrandsdalen. Silty sediments of 0.5–1 m depth above marine clay deposits at Romerike, more than 200 km south of the start point of the outburst flood indicate flood of enormous dimensions [6, 19]. Similar to the soils at Øksna, the silty soils at Romerike were poor in nutrient content and cultivation of these silty soils took place in the 1920s and 1930s after trace elements were included for fertilization and appropriate drainage techniques were applied [20]. Documentation of such flood sediments is of importance for prediction of effects of flood caused by dam failure of a hydroelectric plant.

5. Conclusion

The investigations after the flood in central Norway in 1995 showed that flooding of 2–9 days duration caused decreased yields of cereals as spring barley, spring wheat, and oats. As sufficient N-supply after end of the waterlogged period could bring the yields of barley and oats up to normal levels, it is likely that the flooding with cold running water caused delayed development and maturing of the crop and the fertilization could compensate for N leaching during the flood.

Reconstruction of erosion craters and other measures to make the damaged agricultural areas into agricultural production showed large differences between the best and the least successful measures. The restoration measures gave as a mean about 85% of the yield compared with areas without damage or eroded topsoil. Filling erosion craters with loamy fine sand and mixing in limed sewage sludge gave similar yields as the undamaged areas. This treatment significantly increased the amount of readily available P in the soil, while use of another type of sewage sludge had no positive effect on the yields.

Acknowledgments

The investigations after the flood in 1995 were supported by the Norwegian Research Council, Grant No. 110038/110. The study of effects of measures for reconstruction of soils after the flood in the period 1999–2002 was supported Statens landbruksforvaltning, (now named Landbruksdirektoratet). Harvesting of the plots was performed by Norsk landbruksrådgiving, Innlandet. The farmers Gunnar Sætersmoen and Helge Øverseth at Øksna are thanked for good cooperation.